๐ง๐ Ultimate Startup Growth Playbook, Using AI to Automate Operations | Sam Ross, Numeral
Why retention trumps growth, lessons from the early days of Facebook ads, adding $1B of GMV at Airbnb, how to use pockets of demand to find business ideas, and raising a Series A in four days
This conversation with Sam Ross at Numeral is a master class on all things growth at the zero to one stage.
We talk early growth lessons from adding $1B of GMV at Airbnb, plus stories from being one of the largest, earliest Facebook advertisers.
Sam also explains how to work backwards from pockets of strong customer demand to find business ideas, how to establish early social proof around your product, why retention trumps growth, his hot take on NOT hiring a career growth person as your first growth hire, how Sam raised a Series A in four days, why sales tax is so complicated, and all the ways Numeralโs using AI to automate and make it easy.
Thanks to Susan Liu and Tripp Jones @ Uncork Capital, Chetan Puttagunta @ Benchmark, Lenny Rachitsky, and Nate Matherson for their help brainstorming topics for Sam.
Please let me know what you think of this one after you listen / read through!
๐ Stream on Spotify and Apple
Timestamps to jump in:
3:03 Why sales tax is so complicated
8:06 Running crazy Facebook ads in 2013
11:58 Why you need to be aggressive on new growth channels
16:55 How strong retention unlocks massive businesses
18:24 Using pockets of demand to find business ideas
21:34 Balancing performance vs brand marketing
25:45 How to build a brand from scratch
29:11 When cold outbound actually works
36:18 Building early social proof around your product
43:33 Donโt hire career growth people for growth roles
49:31 Lessons building a jewelry business doing $30m in revenue
58:44 How the 2018 Wayfair v South Dakota decision led to Numeral
1:05:04 Hacking an early product together with spreadsheets
1:07:32 Automating the product
1:15:26 What happens if you donโt pay sales tax
1:20:41 How Numeral uses AI and LLMs internally
1:26:41 How to compete against non-technical incumbents
1:32:57 Why they raised VC for Numeral
1:38:19 Raising a Series A in four days
1:45:43 How big can a sales tax company really be?
1:48:55 Creating a better global tax system
1:54:31 How San Francisco is losing its soul
Referenced:
Sam on X / Twitter and LinkedIn
๐ Find on YouTube, Spotify, and Apple
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Sam, welcome to the show.
Sam Ross:
Thanks for having me.
Turner Novak:
Yeah, this will be exciting. For somebody who's not familiar with Numeral, actually listeners of the show are actually probably familiar with Numeral, but can you just real quick for somebody who's not a regular listener, just explain real quick what it is?
Sam Ross:
Yeah, so Numeral helps businesses in particular SaaS and e-commerce businesses with their sales tax compliance. So if you're a business as you get to scale, you end up having to go and collect sales tax on your orders in different states in the US. So if you sell software into Texas, you need to make sure you're collecting tax on those software orders and putting that in your invoices and then actually going to each individual state and remitting that money. And so we try to make that process as simple as possible and just kind of obfuscate away all of that complexity.
Turner Novak:
Complexity, is it pretty complex? It sounds like it's one of those things that could be super straightforward but also sounds like one of those things that could be extremely time-consuming and easy to mess up.
Sam Ross:
What makes it hard is a couple things. One is that it's like no one wants to think about it or deal with it, this is just compliance, it's not strategically important. So every minute you spend on it as a founder or even CFO, it just feels like a waste of time. And so I think that always makes it even more painful to deal with. But it is complex because it's a decentralized system in the sense that this is not governed by the IRS or some centralized authority. You're actually dealing with each individual state has their own set of laws and within that individual districts can have their own laws. So for example, software is not taxable in the state of Illinois, but in Chicago it actually is. So you have to go and figure out at individual district level what's going on. There's 13,000 different tax jurisdictions, so.
Turner Novak:
Whoa, this is US or global?
Sam Ross:
This is just US alone. So yeah, global is its own set of complexities, I'm happy to get into and we do help folks with,
Turner Novak:
et's talk about that later. That sounds too complicated for right now.
Sam Ross:
Yeah, before I make it too insane. And there's a street in Colorado, it's always an example where I think the west side of the street has one tax rate and the east side of the street has a different tax rate because they're in different school districts. And so we actually, one thing we'll end up having to do is when we get an address, we don't just look at the city, we actually put it into a GIS map and figure out what's that exact rate for that GIS coordinate.
Turner Novak:
So if it's an odd number or even number, it's on a different side of the streets, you have to know and really map out exactly what jurisdiction it's at.
Sam Ross:
Yeah, because we actually have a map of every special district in the US and those actual tax rates. And then on top of that, different products are taxed differently. So here in California, if your water's sparkling, it's taxable, but if the bubbles are not there it's tax-exempt.
Turner Novak:
What? I didn't know that was a thing.
Sam Ross:
Or if you're selling something like an Olipop, it's a soda in some states, but other states because it has 3% juice, it's taxable as juice and the carbonation can come into play. So there's all this crazy complexity and most businesses you don't have a sales tax expert in house. And also frankly, you don't want to have to hire that, that's a waste of energy on your side and that's not something that's strategically important. So really the idea of starting Numeral has been to just always how do we make it so that anyone in your company with zero expertise here can just delegate away all the sales tax knowledge and we're able to obfuscate that complexity for you.
Turner Novak:
Yeah, it's one of those things, it's basically all rules-based, right? So it's like if you build the software that just encodes everything, it's just like, here's the thing, feed it in, and it probably just tells you maybe I'm simplifying this a lot, but.
Sam Ross:
Yeah, no, that's right. It's pretty deterministic in the sense that there's not a lot of judgment calls in. So there's some examples we like to use where there is a little judgment, I think we were talking about edible underwear, where...
Turner Novak:
What?
Sam Ross:
Is that taxable as candy or is that taxable as lingerie?
Turner Novak:
Yeah.
Sam Ross:
So there's cases where things are tough. We have a lawyer on staff, our head of tax research and he'll read case laws and things like that to figure out some of these nuanced cases. But most of the time this thing, it's not like when you have say your income taxes, there's a little bit more judgment going in there and maybe more discretion involved. Whereas with sales tax it's more black or white, it's very complex, but most of the time there's the right way to do things and the wrong way to do things.
Turner Novak:
And I know you've kind of gotten into this because you've started and scaled a bunch of different companies and brands over the years, and I think one of the big things I want to hit on this conversation is some of the different products that you've built over time, but a big component of it has been growth and growing them, and I know we're going to... This episode, we've kind of... I actually haven't thought of the exact title yet, so we'll see how this plays out in post-production, but some sort of zero to one growth playbook, I think can be a good chunk of what we're going to talk about. So I think just generally speaking, when I talked to people when I was like, hey, I'm having Sam on the podcast, what should we talk about? What is he good at? Basically everyone was like, oh, he's really good at growth, you should have him talk about growth. How did you first get into just thinking about growth? Was there a certain moment where you're like, oh, this is pretty interesting?
Sam Ross:
Yeah, so I think it started, my first tech job was this company Teespring. It was this very hyped company with Sam Altman and Keith Raboy on the board. I think at one point we were the number one Stripe customer in terms of payment volume going through the system. This is very, very-
Turner Novak:
Oh, really?
Sam Ross:
Very, very early on.
Turner Novak:
This was like 13? 2013 or something?
Sam Ross:
Yeah, something like that. Spoken in the same sentence as Stripe, Sam Altman came to our offices like, this is going to be a 10 billion company, no question. And so all I have to say is there's a lot of intensity. And while I was there, actually Facebook ads came out, newsfeed ads, and this is a time where people probably don't remember Facebook stock was not doing so well when it went public. There's articles, Procter and Gamble is cutting all their Facebook spend, they say it doesn't work. But actually they were releasing all these intense things you could do on the targeting side. And so that's what our sellers in the Teespring platform were doing. And actually, one of the first things I did is I actually went and started running Facebook ads to sell T-shirts just as experiments to learn. And that really kicked off I would say a lot of my background in thinking about growth.
Turner Novak:
And I know we wanted to talk a little bit about some of those things that you did, like what are some of the examples of ways that you would do Teespring ads?
Sam Ross:
So I think back then it was a little bit more wild west in terms of Facebook sharing their data, this is well before I think-
Turner Novak:
The clamp down.
Sam Ross:
... the Trump winning the election and then media backlash and the clamp down coming from all that.
Turner Novak:
Yeah.
Sam Ross:
So you could do a lot more things in terms of targeting. So one thing for example we would do is you could actually make searches on Facebook and scrape. So we had make a T-shirt, hey, keep calm and drink like a Novak. And make a shirt targeted towards your name. So I will scrapes Facebook for everyone named Novak, get a list of all the Facebook user IDs, upload that as an ad, and then you'd be targeted with the T-shirt with your name and you'd get these metrics that were off the charts, 30% engagement rates, dollar cost per conversion, things like that. And so we were able to do very clever and hyper targeted ads. Eventually what happened was there was a College Humor article that came out one day. It was top 10 creepy... This is College Humor 10 years ago, right?
Turner Novak:
Yeah, I think I remember some of the backlash about the crazy targeted products.
Sam Ross:
Exactly. So it was top 10 creepiest Facebook ads and number one was T-shirts with your name on it. And literally that article came on that morning. By that afternoon all the ads of our seller... Because it was sellers, it wasn't Teespring running the ads, mostly. All the ads had been basically banned. And we found out years later, it went all the way up to Zuckerberg, he was really mad and they actually hard-coded all sorts of things in the algorithm to limit the reach of T-shirt ads in the Facebook platform because I think at that point, something like over 5% of Facebook ad spend was people selling T-shirts on Facebook. Because nobody knew how to...
There was targeting in the algorithm, you could target people that like Chihuahuas or nurses and nobody else was targeting Chihuahua lovers on Facebook, like dog grooming companies hadn't figure out Facebook ads yet, but we could figure it out. Oh cool, I'll put a Chihuahua in a T-shirt, design this shirt, set up an ad in five minutes and I'd make a few thousand bucks from that. And so it was the first way to really... You take that interest graph Facebook and yeah, I think we learned a lot of... There's just a lot of tactics you figured out, I think that was a lot of our learnings on growth is you have these tactics that'll work for a fixed period of time and while they work and while that arbitrage still exists, you really have to go all out exploiting it because probably six months from now it's a market and the arbitrage will be gone.
Turner Novak:
You just think about if you're really thinking with a hyper growth mindset, you have to find these pockets of undervalued channels and just blow them up to the extent of you just have to capitalize on it while it's there and it's going to be gone soon because other people will probably see it and move in.
Sam Ross:
Yeah, exactly. So I think the thing that's tough with growth or honestly it's just marketing, is that the things that are obvious to do, run a Superbowl ad or... Those are probably priced pretty fairly, and if your business is really good, maybe it works for you but it's not going to be exploitable. But you have these periods where Facebook ads are really under priced or when TikTok came out a while ago, you could get tons of organic reach and most people in growth know, okay, this is only going to last for so long, I'm not going to build a durable business promoting my brand on TikTok, but for these few months while I can actually get distribution, you should just go aggressively juice the channel or whatever you want to call it, as much as possible because these opportunities are usually short-lasting.
Turner Novak:
So the opportunity with Facebook ads was people basically thought they didn't work, so there was just no competition, but they were actually very effective?
Sam Ross:
I think people didn't think they worked, they were very effective, and I think it was that you could also do this super hyper-targeted... You could upload a list of one person and target them. So we would always play pranks on each other at the office where it'd be like someone's birthday, we'd make some crazy ad with their face on it and then we'd run it with the big budget, so every time they logged into Facebook... This is even before mobile, it was log into Facebook on their computer and they'd just see ads with their face on it for weeks and it would drive them nuts.
Turner Novak:
That's amazing.
Sam Ross:
We would do all sorts of crazy stuff that was just super targeted.
Turner Novak:
And you would pay a dollar for that, right, or something? Because you're only paying for the impressions that shows up on that person's feed.
Sam Ross:
Yeah, I mean that would be pennies. And there was a game of cat and mouse, so quickly Facebook saw that, like okay, now you have a limit you have to target at least, I don't know, 100 people for the ad. So that was fine, we'd just target a hundred people and then we'd create an exclusion list of 99 people, and so you'd still figure out... There was just kind of this cat and mouse game where they were constantly updating their terms of service and then we would just find different loopholes to continue to target.
So yeah, I mean with the shirts with people's names on it, I remember I could only do it so much and I had a full-time job. I had some friends who were unemployed or half-employed and I taught them how to target people by name, so I remember they had a weekend where they rented an Airbnb, they just bought a bunch of stuff to party, they stayed up all weekend and just launching these Irish shirts with different Irish names on it, keep calm and drink O'Shaughan, or whatever. And I think they made 70 grand in a weekend just because the opportunity was just so easy right then. We used to call it free money, you just put the shirt up, do the thing, you come back the next day and you've made a bunch of money.
And so these things... In retrospect, when Facebook started cracking down and banning it, everyone's like, man, I should have... Why did I sleep? Or why didn't I quit my job and just do this full time? I could have made millions of dollars. And so I think that's often what you learn is sometimes these opportunities, they're only going to last so long. And so you can always do things the most efficient or careful way, but if it's working, you just really got to go aggressive and double down because that arbitrage will be gone soon and you really want to capitalize on it.
Turner Novak:
Yeah, I had Austin Rief, the founder of Morning Brew, I don't know... The newsletter. He was telling me, he thinks that they were the first advertiser on Instagram stories. I don't know if anyone listening is able to fact check that.
Sam Ross:
Yes.
Turner Novak:
But I think we would both actually love to know if they were, because he's convinced that they were. And when it first came out, he told me they didn't sleep for three days, they just were constantly optimizing and running Instagram ads. And it was a couple of weeks where they were getting 1 cent per converted subscriber. So he said it was in the history of Morning Brew, it was the best two week, three week period because they could just acquire someone who's worth 10 or 20 bucks for a penny. And so they were just constantly all in on this thing, just trying to juice it and get as much as they could, because he knew that it probably wouldn't last forever.
Sam Ross:
Exactly, yeah. And these newsletters too, even the value of subscribers goes down over time because everyone's subscribed to 100 newsletters And so maybe you'd value a subscriber at whatever, $5 a couple of years ago, now it's valued at $1.50, whatever it is. And so yeah, if you're first to market in some of these channels, it's valued a lot. I think what I've had to learn is I was really good at these different growth hacks and the first real business I built as an entrepreneur was I was selling these... We had cheap jewelry I'd buy in China for 50 cents for like $50 and we sold $30 million of jewelry, made a ton of cash, but the problem is I was selling something with no retention and so it wasn't a durable business and eventually had to sell it for pennies on the dollar, whereas I think when you build something that's very...
The match made in heaven is you pair these growth tactics that are going to get you out of the atmosphere, but you pair it with an actual business that's you're able to drive and capture value over a long period and it's durable and that's how you actually build something that's sustaining versus you can maybe get a bunch of growth in a short period and then it ends up not being... Maybe get a bunch of cash overnight, something like a gas app or something like that, but it's actually you really want to have something... If you compare those two, that's how you build a really big business.
Turner Novak:
So it's probably then having just the intellectual honesty of understanding this is a temporary ARB and maybe I can make 100 grand or something versus okay, I have a sustainable retentive product that is differentiated over time and however we want to dissect this competitive advantage and the moat that they built or whatever, and just being honest about what the two different lenses of it are and just knowing which one you're doing.
Sam Ross:
I think that's right. I mean, again, no disrespect to doing... I've had all sorts of crazy hustles that made a lot of cash in short periods, so I think it depends, everyone has different goals. But I think in retrospect you saw if I had seen that Facebook ads are working well, then I could go... And again, we were doing different targeting based on the interest map there, meaning I could target nurses, nurse shirts would sell so well.
Turner Novak:
Really?
Sam Ross:
Yeah, I don't know, it's a huge demographic, they have disposable income, it's one of the biggest categories, it's all these things about being pride, like I save people's lives every day, these kind of corny things. You probably see less in San Francisco, more maybe in other parts of the country of these type of shirts. But in retrospect where you shouldn't is like, wow, nurses is this amazing market, what could I build there? And I actually had a friend, he built a company, it was about staffing for nurses and actually built a durable VC backed business for travel nurses staffing.
Turner Novak:
Which one is this?
Sam Ross:
Man, I'm forgetting the name right now.
Turner Novak:
There's one called Trusted Health.
Sam Ross:
Yeah. And I had worked with him before, he wasn't selling, he was... Oh man, I'm blanking on the name, I think Hired or one of these other companies, but we knew that nurses... He had a Facebook ad background, that's actually where he came from and he knew that nurses just were a really active audience on Facebook you could target and sell them stuff. And so I think again, it's pairing that insight of not just on the product side, but is there almost, can I really capture that demand? And so sometimes you find, wow, there's really this demand pool of this market that I can really reach. And so you maybe want to backwards engineer like, okay, what are the types of services I can sell them? And what's the stuff that's going to be the most durable in value capturing business there?
Turner Novak:
Yeah, to the point of nurses, it's like they're probably, they skew relatively higher income just on average. They're probably maybe younger ish, so they have new disposable income versus what they used to have. I don't know, it's an interesting category.
Sam Ross:
Or actually, FIGS is a great example of that, right?
Turner Novak:
Oh yeah.
Sam Ross:
So this is before FIGS... FIGS, they sell luxury scrubs to nurses.
Turner Novak:
It was an insane business. It's like a public company now, isn't it?
Sam Ross:
Yeah, it's a public company.
Turner Novak:
Yeah.
Sam Ross:
And obviously I could have thought about that, I've sold tons of shirts to nurses, instead of selling cheap T-shirts that are printed on demand, I should actually just go and get scrubs built.
Turner Novak:
Yeah, it's like their uniforms, they'll buy it to do their job, talk about retentive products.
Sam Ross:
Exactly. And so there's all these opportunities that you probably... If you squinted hard enough and weren't distracted by the short-term cash of selling T-shirts, you probably could have found, again, these bigger business models that you could sell there. Yeah, there's other ones like truck drivers were a huge demographic. I think broadly these ad platforms tend to work well for, I think... And it depends on which platform, but it's like folks with disposable income who are maybe... Maybe not... Their jobs have enough free time where they're browsing on social media, nurses are busy, but maybe they have in-between time, things like that.
Turner Novak:
Yep. Interesting. So one of the questions I wanted to ask you is what do most people get wrong when they just approach growth? We maybe hit on it a little bit, but anything else you'd want to add on that?
Sam Ross:
Good question. What do people get wrong? I mean I think there's a couple things. One thing I see, my background's more on consumer growth, having worked at Teespring, at Airbnb, and having done my own direct to consumer businesses, so I'm still learning around B2B, but I just see so much boring B2B stuff. I'll go on my LinkedIn feed and I see an... It's like download our checklist of seven KPIs and is anyone actually downloading a checklist of KPIs? That seems insane to me. And I think the approach we've taken is much more, again, stealing what I've learned on the consumer side is like, okay, actually what works well is having some authority and getting... We get influencers to go post about us or using content made by someone with some authority, or having content that's just more engaging and interesting.
Because there's this old term Facebook they used to use, it's called it makes stuff thumb stopping.
Turner Novak:
Thumb stopping, okay.
Sam Ross:
Which is so corny, but it's actually really true. And we're doom... Well, not all, but a lot of us, like I am doom scrolling on my phone and if I immediately identify it's an ad in two seconds and it's boring, I'm just going to skip past it. But if something is actually is engaging and interesting, I think it works better. I think the thing that that's hard most times to reconcile is a lot of folks that are really in performance marketing, they'll tell you, ugly ads work better. And because it's just like it breaks the paradigm, you stop, you look at it. And so usually from a metrics-driven thing, at least you'll get higher click-through rates and things like that with ugly ads.
The challenge obviously with that, I think the mistake you can make there is you're only measuring success metrics. You're not measuring the people who saw that and now are totally turned off by your brand because like, oh, it's just some cheap ugly thing. And so I think it's always a balance, you want to look at the metrics, you want to make things engaging, but you also need to be very mindful of what are you not measuring? Are you cheapening things? Things like that. So I think that's another mistake you can make, too far to the extreme where performance marketers will just think that brand people are total idiots and you can't measure... They're not measuring anything. And I think it's more like you have to have this sort of synthesis between being qualitative and just, if I looked at this, is this going to convince me to buy this thing? Or is this going to convince me to trust this thing? But at the same time to take a data informed approach and make sure what you're doing is attention grabbing and things like that.
Turner Novak:
Yeah, because kind of like when your performance marketing, it's also brand marketing, people are still seeing the product and seeing the brand, so hyper-targeted, but they're still going to have the memory of how it made them feel and how they think about it in the future.
Sam Ross:
Yeah, exactly. So I think if someone that I think has always done fairly well at this is someone like a Hims where they might have an ad that's just... It'd be a side-by-side, a bald head and a hairy head and it's like this eye catching thing and catches your attention and it looks like maybe something like a drop ship or whatever would run. But then you click through and their website is very intentionally designed. Before they even launched the company I had heard they spent I think $300,000 on photography.
Turner Novak:
Really?
Sam Ross:
And really, I think again their thing was like, how do we take this thing that's shameful and elevate it? And so I think again, the people who are best at marketing in 2025 are really able to still marry the best of whatever brand marketing things that are hard to measure and really making things cool and desirable, with also this sort of more data informed approach.
Turner Novak:
So one bigger question then maybe stepping back a little bit, if I'm starting a product or brand or company, whether it's like SaaS consumer, just generally speaking, how should I think about maybe going from zero to one on my brand, my marketing, my go to market? What kind of things would you, if you were advising me, just tell me to think about an approach at a high level? And then maybe we can get a little more tactical.
Sam Ross:
Yeah, I think obviously the type of business matters. So if you're a consumer brand and you think... Just for example, if you're a direct consumer goods brand, you're going to need to... Facebook ads are going to have to work for you or else this business is never going to work.
Turner Novak:
Really? Don't people say don't rely on paid marketing? That's another...
Sam Ross:
Yeah, they say that, but I think pretty much every single scaled up brand that's selling on Shopify, like direct to consumer that I've seen is spending... They're typically their number one channel is Facebook ads.
Turner Novak:
Fair.
Sam Ross:
Or something similar. And so one thing I've done in the past is I would actually create it. So you want to actually figure out what are my economics here? Are they going to pencil out? Because if the economics don't pencil out, it's never going to work. And so in those cases, one thing I would do, I've done before is I would actually go and create a fake Shopify store with a renderings of what the thing I want to sell, I run ads to it, I collect the credit card payment, I put a temporary hold, I don't actually ever charge the people, but you want to see are people going to put money down? And it's not that you need to get the economics down to an exact degree, but you can get a sense, are people willing to actually put money down?
Because If you just go survey friends like, hey, would you buy this? Yeah, sure, yeah. They want to be nice to you, they're not going to be real with you. People are just too agreeable. And so I think if it's something that's consumer, you can actually go and do this testing, spend a couple thousand bucks. And again, I think people will get confused when they do this and they'll go like, oh, I had this good click-through rate or something, no, if people actually put their credit card down, that's what will tell you. And it's hard to always... These things don't work in every case, but I think for less expensive goods, something like that can work. So trying to get a sense of the demand ahead of time.
Similarly, if your thing is, on an analogy there, maybe on the kind of B2B side is do the same thing maybe with cold outreach. And so go... One thing we did really early on was I first started with friends, but then after that quickly I was just cold emailing Shopify brands all over the place and was getting a shockingly high hit rate of folks who were like, yes, I hate the solution I have, or I don't even have a solution yet, but I know I've been needing to do this. And immediately like, wow, if I've only emailed 500 people and I've gotten this many response rate, you can kind of pencil out, actually even just doing cold email loans is going to get me pretty far. And our first, I think million and a half or so of ARR was just purely on cold email, we didn't have any ads or anything like that. And so I think just going and just try to get people who are not your friends to tell you that they would actually go and buy... And get a sense of the demand there I think is really important.
Turner Novak:
I think Ramp famously, I hope this is public, maybe I'm disclosing something, but in the early days, and I think still 90% of their revenue or 80% or probably it's not the exact number, but it's a much bigger number than you would think of their revenue comes from literally just cold outboun.d just getting in front of people saying, Hey, here's Ramp, here's the product. Do you want to use it? Blah, blah, blah. Which is kind of crazy that it's that high.
Sam Ross:
Yeah, they're like the pinnacle example for doing... I think a company that's super focused, just amazing execution at B2B growth, right? Yes, they have all these ads, we're going save you money, or we have Bill Pay. But honestly, my sense is the real reason you had Brex already established in a good company and good at marketing and all this, but the reason Ramp's been able to get such intense growth is I think they really valued building a really high quality growth team there, there's really some amazing talent they've had, and a number of them have started successful startups even. And I think by making it growth like a first-class citizen, and there's a lot of investment in terms of the infrastructure they built, they were able just to out execute. Because at the end of the day, you still need a credit card, which is really a commodity product. And yeah, cold out, it's surprisingly cold outbound just worked really well for them, and I think they were doing it and they had a lot of tactics there under the hood that made their stuff work better than everyone else.
Turner Novak:
Yeah, because I think it's one of those things, if you have this really big top of funnel brand and then you're just always aware of it, and then someone's like, hey, convert right now, or like, hey, here's our product again, I'm asking you to make a commitment to it, or whatever. The specific targeted sniper shot, cold outreach, it's actually more effective than just going in without them even being aware of your product.
Sam Ross:
Yeah, totally. I think the cold outbound, you see it works so much better once folks have been primed already with other touch points. And these things are always really hard to measure. I did some insane work on multi touch attribution at Airbnb and pulled my hair out and was like, "Okay."
Turner Novak:
Oh yeah, really? Oh, yeah? So what'd you learn there or what was the study that you did?
Sam Ross:
Oh, man. I mean we had crazy stuff where Facebook was sending me like... This is, again, before Facebook clamped down on stuff, but they had this program called Atlas where they'd bought what's called a DSP and basically Facebook had tracking data. Facebook puts their pixel in every company's website and so they could track you around the internet on every website you're visiting, and they were able to map that cross device. They had this crazy level data they were giving us in a data room. There was some security here, but log level data, people's traffic on...
And they had a team of, I think, eight data scientists and we had a data scientist on my side. These are folks with PhDs from Harvard in econ, study under Larry Summers, just like GS level people. And we used to take all this, we tried to build these multi-attribution models and you just end up, you maybe build something that looks good for on the historicals. And then as soon as you start optimizing towards those metrics... I can't remember the name of that principle, but when you pick a metric and you optimize towards it, it makes the model not work anymore.
So they would say like, "Oh, you're super undervaluing retargeting and making this up. So let's put a bunch of budget into retargeting now." And as soon as I do that, you think it looks really well. It looks really good, and then you run some study again. And no, that's not the right way to think of it because you're optimizing on some correlation, not something that's causal.
But in short, basically what I'm saying is I think a lot of these things are pretty intractable and it's really hard to know. Someone has so many touch points and half of them are just going to be conversations with their friends, and it can be as subtle as some people think like, okay, I'm just going to only target the buyer of my product. So in my case, "Hey, let's only target say the CFO of a startup."
And it's like, actually you really want also everyone surrounding that person to know about Numeral so that when they go and buying Numeral, it's not like them going on a whim to make that decision. Everyone else says, "Oh yeah, I've heard of that before. Oh yeah, I think that that's probably pretty good." And it just makes the buying decision so easy. So if no one's getting fired for adding Ramp to their startup because everybody has heard of Ramp and there's some trust there. Versus if that buyer had only gotten hyper-targeted from some other credit card startup, but they don't have that authorities.
Turner Novak:
And some money too. It's like, are you trusting your money with this random startup that no one's heard of versus oh, they run Super Bowl ads.
Sam Ross:
Right. Yeah, exactly.
Turner Novak:
Or they work with Saquon Barkley, right? How would you not trust them?
Sam Ross:
Yeah. So I think this is what makes this measurement part of marketing is super hard. And people have tried all this stuff and I think there's ways you can do some stuff. It's not that you want to ignore the data, you kind of look at stuff bottoms up and you make a best guess. But a lot of it, for me at least, it's just trying to run just a lot of back of the envelope math, trying to have a good intuition around it. I think I'm fortunate that I have a lot of tacit knowledge from being in different places. So Airbnb for example, there's this thing on Google where Google, they're very exploitative in that you have to pay to be on your own keyword, right?
Turner Novak:
Oh, yeah.
Sam Ross:
So at Airbnb we were spending... Oh man, the budget of one year I think on Google Ads in like a billion dollars. So maybe we're spending-
Turner Novak:
What?
Sam Ross:
You were spending 50... maybe it's cut down now. We were spending something like 50 million bucks a year just to bid on the keyword Airbnb. So Booking.com didn't show up above it. We ran all these crazy experiments, like difference in difference, studies where we had these data scientists trying to run stuff at geolocation and to figure out how incremental is bidding on our own keyword.
I get a sense... I can't remember the exact number. It something like maybe 15% of it's incremental, but you see enough of these studies over time and you get a sense, okay, if I'm bidding on my own keyword, I should probably only trust that something like 15% of those conversions are incremental and the rest of them are just kind noise.
Turner Novak:
Would've converted anyways.
Sam Ross:
Would've converted anyways, right? And so-
Turner Novak:
They would've searched Airbnb and be like, Booking.com, that's not what I need it. I'll keep scrolling up. Oh, Airbnb there it is.
Sam Ross:
Exactly.
Turner Novak:
I'll take the second or third.
Sam Ross:
Exactly, but maybe 15% of those people, they would've just clicked the Booking.com thing, right? And so I think a lot of this stuff is... I can't run that experiment that Airbnb did. I could never power that study. And it's also, it's a pain in the butt to run the analysis. And so a lot of the time it's like you get an intuition around how effective something is and you can kind of discount it or give it extra credit based on intuition.
Turner Novak:
It seems like, to your point, intuition, it seems like a social proof and that kind of stuff is a pretty good strategy for generally building trust around this stuff. How would you recommend doing that? If I'm a non-social-proofed product, if I'm like, man, I need to build some social proof, how would you recommend approaching it?
Sam Ross:
So I mean what we've done... Lisa's worked for us at Numeral. We started off in the e-commerce space. And so I think one thing early on is we knew this is a strategy we wanted to take. And so we put a lot of effort into trying to figure out good angel investors that we knew would talk about us in the future. So for example, there's this guy, the founder of... Moiz Ali, he started-
Turner Novak:
Oh yeah, Native, the deodorant.
Sam Ross:
... Native deodorant. I think he owned the thing pretty much outright, sold to a hundred million to see Procter & Gamble. But we knew we wanted to get him not just because he was a good name, but he's someone who I think is highly trusted. A lot of folks in least the e-commerce space, they'll shill anything and they're named. Their advocacy doesn't mean much, but he's someone that's very disagreeable, will be quick to call out people who he thinks they are scams, things like that.
And so we knew if we had Moiz eventually tweet about us at some point... And similarly we had another influencer, Shaan Puri, My First Million podcast. We knew if we had them on our side, that would mean a lot. And Shaan, he has his own brand on the e-commerce side that we service. And so again, the authenticities there, you could see you can talk about.
Turner Novak:
I actually use this.
Sam Ross:
I actually use this. And so I think for us, it's not always easy to this, but if you can get angels who ideally actually use your product. And again, it's not necessarily Angels that have had a lot of reach per se, but it's folks who are just, people really trust their authority and actually don't talk about other products they use very much. Then the value is really high there.
And so we've been able to leverage that. I think that's just quickly built us a lot of authority really quickly out of the game. People hadn't heard of us before, but a lot of time when you're busy it's hard to evaluate different options in a space. And so often if someone you trust is using that product, that's a quick meta rationality to go and delegate your searching. Especially in a space like sales tax, nobody wants to go check around four or five, six options. That's not a fun thing to take a demo on.
Turner Novak:
Yeah, I'm taking a day to figure out what sales tax provider I'm going to use.
Sam Ross:
Especially when we talk to founders, often they come or they call and they're like, "Look, I don't want to look around here. Just solve this problem for me. Make me think about it as little as possible." And so if they've already heard someone they trust is using it, that makes the buying decision much quicker versus maybe it's a fun product to use and I want to really just demo it for fun. Then you have to spend a lot more time just making a magical experience with the product or something like that.
Turner Novak:
Yeah. How do you get the first person on board? How did you get Moiz? That sounds like that was a strategic, we got to convince this guy. How did you get him?
Sam Ross:
For sure. So the first one was actually crazy, was I tweeted... Before we even started the business, we were just thinking about it like, "Hey, I'm thinking about doing something around sales tax." I had a hundred followers, or I had 200 followers on Twitter. Like, "Would love to talk to some friends." And Shaan Puri somehow saw that tweet just DM'd me. He is like, "Yes, this is an obvious idea." And he was I think our first or second customer. So that I was just-
Turner Novak:
Wow.
Sam Ross:
Honestly, that's not good advice. I just got lucky there.
Turner Novak:
It's that Twitter algorithm or maybe someone sent it to him.
Sam Ross:
No, I think he's just on Twitter all day. I don't know. Yeah, I have no idea because I had 200 followers and maybe someone sent it to him. But he DM'd and we started. He was an Alpha, he was our first second customer. And then before we did YC, your terms changed. So we made sure we got him a good investment before we even did YC-
Turner Novak:
Oh, cool.
Sam Ross:
... so he'd feel good. He had a big enough stake to care. And then someone like Moiz... I just went through. I knew someone I wanted to target so trying to get an intro to any VC. You just go through your network. Who's my mutuals? I found the right people. The other thing too is you have this density of talent. So he personally knows Shaan. I think someone else made the intro, our friend Adam, the founder of a Postscript. But he is like, "I think"... Again, he knew that Shaan was a customer and an investor and fan of ours and so that made it easier. So the other thing too with getting people to advocate is there's this weird phenomenon where the first couple are harder, but then as you get more and more, everyone else, the risk for someone to go and shill your product... or not shill, but to go talk about your product.
Turner Novak:
Yeah, you're lending trust and lending credibility.
Sam Ross:
Exactly. And so I think even you saw something like FTX, they actually did a really good job at going and getting people to promote them. And maybe Larry David wouldn't have gone and done it out of the gate, but already Tom Brady, Steph Curry, they're already doing it. And so every incremental person is actually probably easier to get because it's less on the line for them. Because if something goes wrong or the startup ends of not succeeding in something, no one's going to get... people were not that mad at Larry David.
If Larry David had been the only person to learn about FTX, maybe it would've been a bigger of reputation hit, but because there was 20 other big influencers talking about FTX, it kind of de-risked it for them. And also he didn't necessarily have to do his due diligence in the same way.
Turner Novak:
Yeah, that's fair. Yeah, and actually I think, like we mentioned, people who are familiar with Numeral, you guys have sponsored the podcast in the past, and I just know Nate on your team, your head of growth, pretty well, and that just helped me get more comfortable with the product of knowing. I mean, Nate's a really sharp guy. He's not going to join this company and grow it if he doesn't understand it. And then also I know a bunch of the people that are customers and use the product and like it. So it's like okay, it makes it a little bit easier.
Sam Ross:
Yeah, I think there's this thing in startups where I'd heard it, but I didn't really understand it until I started of momentum really mattering. And so you have this thing where it's like we were doing pretty well and Nate was actually an angel investor of ours and he knew we were doing well. So when he was shutting down his company or selling off his company and looking for something new to do, he was excited to join because we were doing well. So again, the momentum of doing well gets better talent.
And then by getting better talent, Nate's been a force multiplier on our team on really just incredible growth. So then he grows our business more, our business grows faster. You get additional talent or additional investment or better logo investors and these things all kind of compound onto each other. And so yeah, the cold start is always really hard, but I think as you get into the atmosphere, this is something that I'd seen a little bit before, places like Airbnb, but I can really feel it in my bones now.
Turner Novak:
So I know you have this hot take on not hiring growth, career growth people to do growth. We just talked about Nate who's like the head of growth at Numeral. So why shouldn't you hire... It sounds like hire a growth person, hire somebody who knows growth. Why not?
Sam Ross:
Yeah, good question. So I have a lot of friends that are career growth people, so no disrespect.
Turner Novak:
Yeah, no offense. All right.
Sam Ross:
There are a lot of talented folks out there. I think what I've found is with broad, whatever you want to call growth... I mean honestly, growth is really mostly just marketing, but it's trying to make it higher status so engineers actually respect you, right? Because engineers think that marketers are just some lowbrow shillers or whatever.
I went to go hire a head of growth and I started looking like, who are the people I know that are best? And everyone that was good that I wanted to hire, they were actually just entrepreneurs. Because if you're good enough and confident that you can actually go and drive demand for something really well, you'll end up... Why are you going to get paid some fixed salary? Growth people don't get paid salespeople on a commission. They get paid some fixed salary.
So when I was at Airbnb, I figured out how to drive an incremental billion dollars of GMV. I was really proud. We figured some stuff. We did a bunch of stuff, built a bunch of stuff on the Facebook API and we're like, "Holy crap, we just crushed it." And then they're like, "Cool, you can get a 3% raise at the end of this." And I was like, "This is insane. They should be paying me $10 million or something or more."
And so that was my impetus, one of my impetuses for leaving is I was like, okay, I'm confident in my abilities. I'm just going to make way more on my own. And everyone else I know is good at growth has ended up going and doing that. And I think when you go from playing with house money to playing with your own money, you just learn a lot more because every decision really matters and any piece of pain from mistakes, you really feel it.
And so when I went to go search, I talked to a lot of folks and I realized after certain months, okay, I actually have to just find another founder who is listening for something new. And I got lucky where Nate, again was one of our angels. He happened to just message me.
Turner Novak:
Oh, nice.
Sam Ross:
I remember I responded to his email within four minutes and we gave him a job offer within two days.
Turner Novak:
That's amazing.
Sam Ross:
So sometimes you have to go really fast there. But yeah, I think anyone that's at least, if they're either founder or just they've done some type of growth stuff with their own money and on their own without a system where they just come in, plug and play, you just learn a lot more.
Now there's obviously exceptions to the rule. There are people, I have friends who are incredibly talented, but I think broadly too, growth and marketing is not always attracting the highest level of talent in the sense that all the super ambitious Stanford kids, they want to go out and be product managers or VCs or something like maybe a little higher status or pays better. And I think you end up, people a lot of the time they end up accidentally in growth. And so maybe on average the average career growth person is just not that same level of elite talent.
Turner Novak:
I think it's interesting with Nate because he's been in sort of growth and marketing. His startups that he's built have been around like SEO and helping you grow your company. That was the whole, his product for the past couple of years. So an interesting hire.
Sam Ross:
Yeah, I think it's non-traditional. Yeah, he'd done some SEO stuff. He hadn't done other... I knew that he had a big following on LinkedIn and some following on Twitter. And he was just like within the YC community, there's this internal social network called Bookface, and he's an iconic poster there. People know him.
Turner Novak:
Is he really? Okay.
Sam Ross:
Yeah. He's like-
Turner Novak:
I'm not on it so I wouldn't know.
Sam Ross:
Yeah, yeah, he's a legend. He posts all the time, and I mean he's invested in over 150 YC companies from his own personal fund. I knew he was super networked and even just showing that, hey, you can build a personal brand and personal followers, that's just marketing growth in and of itself. And so it's not like... I think sometimes an engineers, you can run them through a code test. With someone on the growth and marketing side, you could ask him pointed questions around different things, but actually seeing what you've actually accomplished. And he's also built businesses himself as well, and he's gotten big logos and stuff and so I knew.
And again, the skill set you're literally looking for in a growth person is just super high agency, creative thinker and Nate's always got some creative idea. In fact, he sent one of his own company, he'd sent me Hot Sauces he designed with his brand on it and-
Turner Novak:
Oh, nice.
Sam Ross:
a good growth person always you ask them, "Hey, I need to solve this problem," and it's like they can quickly unblock themselves. Nate's always like, "I got a guy for that. I got a girl for that." We came in. Within a month he's like, "I have 14 different content writers for our SEO. I got two editors. I got my designer doing this." He already had a whole roster of folks that he could pull from.
And I think that's what I've seen from anyone that's been doing this type of stuff for a long time, is over time you kind of accrue this network of vendors typically on the growth side. So if someone doesn't have a network of vendors that they want to bring in on day one, it's hard to learn that stuff from scratch because you only know if a vendor is good if you've actually worked with them before.
Turner Novak:
Yeah, that's fair. And speaking about things you've done before, so you talked a little bit about you were at Airbnb, added a billion in GMV. But while you were doing that, you also were launching your own brands and companies on the side. Can you just talk about that era of your life and some of the things that you did?
Sam Ross:
Yeah, yeah. So I mean, when I was at Teespring, as I mentioned, on the nights and weekends I was a prolific seller. Sold a few million dollars of T-shirts, whether it's someone's name on it or yeah, all sorts of funny, funny content on there. And then when I went to Airbnb, I kind of put down and paused for a while and I was actually focused on my job. And I think at a certain point when I was, I don't know. I think when I was told I was going to get a 3% raise, I was like, all right, this is probably not... And the company I think was just becoming a little bit more, we went from this kind of early startup, like high energy place to... Well, Chesky will go on about in terms of being more corporate in the culture. I think hopefully it's changed, gone back to being a little high energy before, but with slowing down.
I realized, okay, I want to escape. And I started running some side hustles. And I had learned from Teespring a lot; again, how to do this targeted advertising. So I actually started selling the same cheap jewelry online. So we had this concept where we had one or two pieces of jewelry and then we'd print thousands of different backgrounds on it. So for example, we started with the demand side. So we see, okay, I wanted to build a business that worked well on different search channels. So we saw people looking for, search for gifts all the time, but they'll look for gift for my piano teacher, gift for my mother-in-law, gift for my cousin's bat mitzvah, whatever it is.
And we were like, "How do we create a product that's hyper targeted towards all of those?" And so we have one piece of jewelry and then we just print thousands of different backgrounds to put on the back, like a paper background in the box to put on the jewelry. So you'd run an ad and it'd be this big colorful thing, I love my bat mitzvah. And they switched for bat mitzvah. And we could just run that on Google or we could put that on Amazon, Etsy, all these different platforms. So-
Turner Novak:
It's the same piece of jewelry?
Sam Ross:
The same. We ended up with a couple, maybe a five or six, but it started just one piece of jewelry, and we just print on demand, these little pieces of paper. So I started off, I was just in my house cutting paper at night after work and putting it in our mailbox. And eventually we had to get a whole manufacturing situation, like a print on demand.
But yes, that was my first foray. We ended up selling $30 million of jewelry over time. But the real impetus for me was I grew up in the Bay Area and then after college I came back for seven years. I needed a break. And so I decided to live abroad for a couple of years, do the whole digital nomad thing in 2017. And so this was a way for me to sustain my life and have something low commitment. There was bootstrap, no investors telling me to come back to San Francisco. I can do it on my own terms.
Turner Novak:
So we talked a little bit about the fake ad testing earlier and getting a consumer brand started. But maybe just taking that fully all the way, if I'm sitting here and I'm somebody who's like, I want to start a brand or maybe start a product or company, how would you recommend approaching that? And maybe it sounds like, maybe how did you do it with the jewelry?
Sam Ross:
Yeah, I mean, consumer's always harder, as you know. As a VC, I think... I don't know, at least in YC, it's only 10% of the companies are consumer these days.
Turner Novak:
Yeah, I'm one of those VCs that does way more consumer than the average would tell you to do.
Sam Ross:
I shouldn't say this, but I think doing the consumer business is way more fun, at least with someone with a marketing kind of background like consumer marketing, you can be a lot more creative and you get a lot more data to work with. And so I think it can be fun. It's just obviously the challenge is the durability. Even some open AI, their retention metrics, they don't look too good or cursor these placings that blow up.
But no, going back to it. I think it depends on what you're selling. If it's your boot... I think if you're doing e-commerce these days, you probably don't want to take too much venture money. Maybe it's bootstrap, you've got some loans or some family or friends. I think the first thing you're looking for, as I mentioned, is you just really want to understand first, if it's a physical good, the economics really matter. And that gets down to crazy details.
Before you even build your product, you should actually figure out your shipping because if you're like 7.9 ounces, your shipping rate's going to be significantly less than if it's 8.0 ounces, or the shape of the box and all this stuff. I think if it's physical good, you really want to map out all of the details of the economics. Now these days you thinking about tariffs and all of that and really make sure that that works.
We had this brand where we were selling paper flowers online, and we found this woman in Michigan and she was making these beautiful bouquets out of paper flowers. These are stunning, but it would take her 12 hours to make a bouquet. And we're like, "If we could do this at scale and productionize it, we'll be able to build something really beautiful." Because flowers obviously, part of the problem is they go bad. So you can kind of sell it. On the inventory side it's hard, but if you could have inventory that doesn't go bad as flowers, you can sell a lot during Mother's Day and times like this.
Turner Novak:
Plus flowers, they're like a three figure, probably a three figure purchase. Someone will drop a hundred bucks on a bouquet.
Sam Ross:
Exactly. And so we did it, but we probably didn't spend enough time thinking about the economics. So we spent a year training up this factory in China and they had all this expertise. It was actually really hard to get that door. We got it to work and just the economics just never penciled out. If I had been able to get my cost of goods down, probably 30, 40%, we probably could have built a business doing I think at least 30 million a year in sales because people really loved it. But we had to charge $300 for this thing, and it just immediately, a lot of people are not going to pay 300 bucks. They're looking for a hundred dollars or less gift for Mother's Day.
And so I think really making sure before you go and do all your investment stuff, even if the product's good, the economics ultimately have to pencil out.
Turner Novak:
I mean, it's probably the same thing on the B2B side too then, is like you want to figure out your go-to market, what a product's going to cost. And I feel like in software, a lot of that now is CAC. It's like how much is it going to cost to sell this thing? Maybe a lot of the AI tools, there's a gross margin component of the cost of building and serving the product and the models to people. Do you think that's fair?
Sam Ross:
Yeah, I think that's fair. I think talking a lot about growth, ironically, I would say growth doesn't matter at all if you don't have good retention. I'd much rather have a business that has 95 plus percent retention or a hundred plus percent net dollar retention that's growing pretty slowly than some business growing super fast. But it's people are churning really quickly.
So again, you went back to this example, but my evolution over time was I started a jewelry business. We sold tons of it. I had years, I made ton of cash, and then by the end it got more competitive. We had all these people ripping us off in China. Ultimately, the economics stopped penciling out. My ad costs got too high, and I sold the business for a fraction of what I could have maybe at the peak.
I saw, I was like, okay, maybe I want something more retentive. We went and built a vitamin gummy company business called Fling Wellness, and we sell vitamin supplements. And we were like, "Okay, why did we build that?" Because vitamins, people take them every day. You can get them on a subscription plan and it should be much more durable. And on a relative basis, it was much more durable. But at the end of year one, still, I don't know, something like 80, 90% of our customers are no longer buying on a monthly basis. And that's just typical for most consumer businesses.
And so I was like, all right, what's the one level up of retention, is something where it's a B2B SaaS. Then within that, we picked up a compliance product where if you decided you don't want to use Numeral... We never do long-term contracts because I think we want to just deliver our goods every day. If someone's not happy, I'm not going to make them stay.
But B, if you decide to stop using numeral, you can't just go and stop doing sales tax. You have to just go and find an alternative provider. Otherwise, you're going to get a bunch of letters in the mail, angry letters or penalties from states and a lien on your business.
And so I think I've gone up the... And so I can grow this business much slower than maybe... It took me a lot longer to hit... It's the same revenue targets as my e-commerce businesses, but it's just so much more durable. I could turn off all my sales and marketing and my business would still grow next year, whereas I turned off all my Facebook ads for... I mean, these e-commerce businesses by this time next year-
Turner Novak:
Plummet to Zero.
Sam Ross:
... it's basically like zero dollar. Yeah, basically zero sales.
Turner Novak:
So then I think this leads to the obvious next question here is just how did you get started with Numeral? You figured out this problem with sales tax, you lived it. How did it all come about?
Sam Ross:
So my first foray into sales tax was much simpler. So when I started 2017, I had business based in California. I had to collect sales tax for my California customers who were buying from me. But I shipped to, I don't know, Ann Arbor, Michigan. I don't have to think about the taxes. I ship it tax-free. It's pretty simple. It's in my local CPA. He takes care of my Californian sales tax for me. It's not something that's really too much of a headache.
And then what happened was at the end of 2018, there was this big supreme court ruling called Wayfair versus South Dakota, where South Dakota sued the furniture brand Wayfair and said, "Even though you don't have any employees in our state or any physical presence, because you're shipping so much inventory into our state and using the roads and everything, you need to be collecting all the sales tax here."
And so there was this landmark 5-4 decision. And basically after that there was a whole set of adjudications in different states. And ultimately businesses like mine had to go and register in all these different states. So each state has what's called thresholds. And so if I sell a hundred thousand dollars into Florida over the last 12 months, then I have to go register and start filing there.
So because our business was already at scale, we had to register in something like 40 different states. And then I started dealing with it personally where I'd get physical mail from all these states, letters from the Department of Revenue, Alabama, West Virginia. I had a case where there was some issue where I think we had missed a couple of filings. And West Virginia, was trying to charge me all these penalties. I'm calling up these folks in West Virginia and trying to beg them to waive a penalty for me and make some other changes. And this is a bootstrap business with two full-time employees. And why am I, as the founder, dealing with this weird esoteric compliance burden and-?
Turner Novak:
Physical forms, calling people on the phone.
Sam Ross:
Exactly. Yeah. Literally, I just get a stack of mail every week from all these states, and every time you get it's totally anxiety-inducing because half of is fine. California sent one to all of our customers last month. It was like, "Breast pumps are exempt for the next year and a half on sales tax." And it's like, "We have no clients selling breast pumps. I don't know why you're getting that mail." But sometimes it's like, "Yeah, we're putting a lien on your business because you miss these filings, so you can't miss it." And so I'd always have stacks of mail that I was just putting off opening.
Turner Novak:
Yeah, because it's a bill. It's just this thing, it's like a government envelope that shows up and you're like, what's inside? How much am I going to owe somebody?
Sam Ross:
Yeah, it's just like every time you open the mailbox and it's a letter from a government tax authority, it's just an anxiety inducing experience. And the incumbent software that we were using, this one called Avalara, I think it was really designed in the early 2000s before you had all this cross state kind of stuff that we're talking about. It was designed for bigger enterprises. It's like a piece of software like SAP or something. It's designed for tax professionals. So, if I'm head of tax at this company or a controller, for someone like me that has no expertise, it was actually really hard to use. The sales people there were used car salesmen trying to scare you into stuff and upsell you on things you don't need. And so, I was like, "This is crazy." And I think, again, I've been doing the... The e-commerce stuff, I did... It was never some long-term vision, I'm going to be selling supplements online. It was sort of something I could do to travel... I could do while I'm traveling. And then when COVID hit, I had to come back from, at that point I think I was in Italy, come back to the US and start working hard. And I'm like, "I was a PM for many years. I miss building software." And so, it seemed kind of like the confluence of having this problem and missing building software.
Yeah, I messaged my old boss at Airbnb, Gustav, who's a partner YC, and I'm like, "This seems like a good idea, but doing a venture-back thing, that's like a life... It's a big commitment for the next 15 years," whatever. It's like, "No, it makes a lot of sense. Go do it." And so, that was the impetus to starting this was again, how do we take this thing where it's just a lot of brain damage if you're a founder in particular to think about and try to take products that are clunky old school and a very narrow thin product experience and create a lot more... Make a much thicker product that really just owns this boring problem end to end. With our goal is that you never really have to actually think about sales tax. You see all these startups where when I was first meeting with investors, like, "What's your engagement metrics? How many times are they logging in?"
And I'm like-
Turner Novak:
It's sales tax.
Sam Ross:
... "Oh no..." I'm like, "This is a negative correlation." The more you log into our product, that's worrisome because ideally you... We have customers who will never log in for months. And I met some woman at a conference and I'm like... I was like all excited to meet her and shake hands. And she's like, "Do we work with you?" I'm like, "Yeah, I was on the sales call. We signed you months ago and you've paid us five figures in fees." And she's like, "Oh..." This woman, Rose, and she's like, "Oh yeah, I think, yeah, you do my sales tax, right?" And then at first I was offended and then after I was like, "Wait a minute, that's exactly what I want," is this just bleeds into the background and you never have to think about this boring compliance problem.
Turner Novak:
Yeah, that's fair. But I could see that the framework, if you read the blog posts about SAS, it's like you want the specific DAU/MAU or DAU/WAU metrics-
Sam Ross:
Yeah.
Turner Novak:
... and they want to be logging in at least 70%, because it means they're logging in every day, and then maybe on the weekend once. So engaging.
Sam Ross:
Yes.
Turner Novak:
I could see how that's the framework you might approach this with. So, when you started though, you talked about building a product, it was just kind of like, "Some spreadsheets." How did you get started actually putting this into motion?
Sam Ross:
Yeah, so we started it before our co-founder, current co-founder, Matt, our CTO, joined. And we were just as a team going and doing things by hand, downloading data in weird funky ways, doing insane workbooks in Excel.
Turner Novak:
You built some spreadsheets that did the things.
Sam Ross:
We had some spreadsheets that had some formulas, but often just even normalizing the data you were having to go and do just a lot of manual work. And then actually going to these state portals, the hardest part often is just the data entry. So, doing a tax filing... Illinois just has this user interface that's designed probably 20 years ago, and every single local... You can't just go and put the tax rate for a local jurisdiction. You have to go and create... It's five or six clicks for each of the hundreds of tax jurisdictions. So, it can take well over an hour to go do a tax filing in a state like Illinois.
Turner Novak:
Wow. That sucks. It's like per file per customer.
Sam Ross:
Exactly. Exactly. So, my co-founder Kevin was like... He was just going and working late and just inputting things cut and pasting. We had my virtual assistant, Michael, who I'd worked with for over a decade, he started with T-shirts for $3 an hour, designing T-shirts. We've upskilled him over the years and he's gotten good at Excel. And so, a lot of it's just training folks like him to go and do a lot of this work, and a lot of learning on the fly. And so, we were able to actually get a first set of customers without any engineering.
And it was kind of funny going into YC where I feel like most of the companies there were you had these 22 year olds from MIT dropout and they're just genius engineers but maybe didn't have a lot of business or go-to-market intuition. And we were really indexed on the opposite where we were very commercially minded and we also knew exactly what we were building. There was no idea maze, and it was just trying to brute force stuff for a long time. And then fortunately now it's all largely mostly automated and you don't have to think about this type of stuff. But, I think, at the start you can often get off the ground without having to necessarily do the big fixed cost thing of building it. It doesn't work for every product, but for some you can.
Turner Novak:
So, you automate that, the state of Illinois's example of five pages clicking through navigating at that point or this point you could... That's-
Sam Ross:
Yeah, so someone like Illinois have you have an electronic filing system. So once we had enough cut clients, you go to them, you say, "Hey, I'm a firm I have, I don't know, 500 clients filing in your state. Can you let me get access to your-"
Turner Novak:
API.
Sam Ross:
... "API?" These things are not... They're SOAP APIs designed from 30 years ago, crazy XML files. But yeah, they'll have some API. Some states don't have that. And so, a lot of it we've had to also just invest in browser automation and we're fortunate now that in 2025 browser automation is a lot more seamless with powered by LLMs than maybe even a few years ago where you're debugging these Selenium scripts every day and it's-
Turner Novak:
Which one are you using right now? Are you allowed to say?
Sam Ross:
Browserbase is some of the infrastructure that we use, and then we would plug in a bunch of different models there as well.
Turner Novak:
Cool. Yeah, we had Paul at Browserbase on the show a couple months ago.
Sam Ross:
Okay, yeah.
Turner Novak:
He's all about browser automation. He's like the browser automation expert.
Sam Ross:
Yeah. I think there's so many workflows out there that I think... It's kind interesting, you can take these... You can almost create an API out of thin air in some ways. And so, we have things where, again, a lot of these states they don't offer APIs, they don't have ways to access the data, but if you go log into a portal and the data is there and it's scrape-able, so it transforms a lot of the things we can do.
Turner Novak:
Yeah. I guess I'm interested in learning how you went from scrappy operational, doing things that don't scale, the spreadsheets, to it sounds like you brought on a CTO, a lot of things are automated. How did you go through that phase?
Sam Ross:
Yeah. I think a couple of things. I think one is once you start having enough traction, again, this is the momentum thing, is Matt was like, "All right, you guys are building a real business here." Like, "This is exciting to join." So we could get... Matt was one of the first engineers at Stripe, one of the first five engineers at Notion, we could get some world-class engineering partner to work on it, versus I think it's times harder to convince people before you have any traction to quit their fancy jobs. At that time he could have joined OpenAI and gotten paid heavy seven figure salary. So, I think, that helped. And, I think, we're also very mindful that you never wanted to do the manual thing for too long. There's been these lessons, like places like Zenefits where I think they scaled really fast before they actually automated all their payroll compliance and they started making tons of mistakes.
Turner Novak:
That's the thing they got in trouble for, right?
Sam Ross:
They got in trouble for... I think there's a couple... I think they got a lot of scrutiny for maybe going around some rules on the legal side. Largely what actually really killed them from my understanding was honestly they just didn't deliver a good product to folks. And so, it all kind of compounded all the bad PR, but what was really happening is they were doing payroll for folks. And again, similar to sales tax, you have to go to the local governments and file payroll taxes on a regular basis. And because they hadn't actually automated these systems or built robust systems, they were doing things by hand, making tons of mistakes on people's payroll taxes, customers were churning and then their reputation got burned and it went from like, "This is a hot company, this is where you should default your payroll to," to actually like, "Don't trust them. Go use Zen payroll," which is now Gusto. "Instead they're going to have a much more reliable thing they've actually built... It just works a lot better."
And so, I think, that's always been something we thought about a lot is your reputation just matters so much for something like this. And so, when we make mistakes, we actually have a guarantee and we'll cut folks checks, we wire people money if the mistake's on us and we'll pay those penalties and interest if there's something went wrong, just because A, the right thing to do, but B, it's also the right business decision.
Turner Novak:
Yeah. I feel like a compliance, security, taxes, these are all things where it's you just want to trust that they're going to not mess it up. And there's probably something you don't want to over-engineer it as a customer or... Just solve it for me and just make sure it's okay.
Sam Ross:
Yeah, exactly. It's really binary in some sense. Just make the problem go away. And the things that we do, I think that it's differentiating or better is try to make... Just reduce the amount of headache and brain damage, and just be as consultative as possible for folks so they really understand what's going on without being pedantic and making it hard to understand.
Turner Novak:
And you've been very focused on just building the infrastructure globally just from the beginning. What was the importance of doing that and how'd you do it?
Sam Ross:
Yeah. To be honest, it wasn't always from the beginning. I think we started off with the US, because the US sales tax is more complex by an order of magnitude than anywhere else, because in US there's 13,000 different tax jurisdictions. And like I said, one side of the block in Colorado is going to have a different tax rate than the other side. Whereas, somewhere... The complexity of the problem is a little bit different internationally. And so, in UK it's a flat 21% V or GST tax rate there. And so, there's not this local problem where you need to understand all the nuanced tax rates, it's a lot simpler. But, what makes it I think a lot harder on the international side is going and getting registered in all these different countries, hooking into all the different portals and places like that, dealing with different language issues, getting money movement, things like that.
So, the nature of the problem looks different. So, when we first started with e-commerce, it was kind of okay to just be US focused. And we had a handful of customers in Canada and we added that pretty quickly. But, when we started selling to software businesses very early on in your life, you're selling globally. And so, as soon as you do a single sale in a country like Saudi Arabia or UAE, you actually need to get the register right away or else you're actually out of compliance and starting to build some tax exposure. And the risk always with sales tax is you did a bunch of orders and you never collected the tax on that. And so, when the authority finally comes to you, you have to then go and pay those taxes out of pocket that you could have passed on to your customer.
Turner Novak:
So, you want to just be collecting it right at the point of sale, at time of delivery-
Sam Ross:
Exactly.
Turner Novak:
... or time of invoice or whatever?
Sam Ross:
Exactly. And so, that's always the risk here. I think it's one of these things where it's like a lot of the time it's like you can go do the order in Saudi Arabia and it's not like the government's going to come and find you tomorrow and regulate you. But, maybe you've been doing it for two years, you forgot about it, and you finally go to get bunned up there and now you have a few million dollars of historical exposure that you waited to deal with, and that's where the pain comes. And then on top of that, there's often be penalties and interests, and audits, and things like that,. But, it's really like that there's all those tax dollars you could have been passing on to customers now that's your responsibility. And we've seen people tens of millions of dollars of exposure that all they had to do was just go get started a few years earlier. And as a founder, you're busy, you have a million things to do, it's a really easy thing to just keep punting on and not thinking about.
Turner Novak:
Yeah. Because, it's probably like... So, if you don't pay your sales tax, I'm assuming you don't go to jail, but it is probably not that bad, but what happens if you just don't pay sales taxes?
Sam Ross:
Yeah. It's a couple of things. So, the case of where you could go to jail is you've been... I've seen folks where they've collected sales tax, but they didn't remit it. They turned on sales tax in their Shopify like, "Oh yeah, I'll do collection." And then they never thought to actually remit it and that's technically fraud.
Turner Novak:
Yeah, I was going to say that's literally fraud.
Sam Ross:
And so, I don't know if I've seen too many people go to prison, but you can... Or you can get felony charges for. That's less common. Yeah. The bigger risk I've seen is I saw a business that someone like a marketplace is particularly affected by this, because you have a... Your revenue's just... Your taxable revenue is really your GMV. And so, if you're running lots of dollars through your marketplace and you're not collecting for a long time, there was a business that had $25 million of tax exposure liability-
Turner Novak:
So, what does that mean? It's like accrued taxes that they have not paid yet or collected?
Sam Ross:
Exactly. It's like if they had gone and registered at the right time... Let's say they were signed across the country, and they should have registered in Michigan two years ago and they didn't. And so, now in the last two years, they should have collected and remitted a million dollars in sales tax and they didn't. Now if Michigan goes and audits them, they'll have to go find a million dollars and pay the state. And so, I met these guys really early on, so I didn't have the expertise to really help them, but they had $25 million of historical tax exposure. So, if the states had all gone... Luckily these states it's, again, a decentralized, so-
Turner Novak:
Yeah, it's not coordinated.like, "Let's all go at once."
Sam Ross:
Yeah. There are some. Sometimes they'll do data sharing. EU would do data sharing, and there's things where I think the Great Lakes region, Minnesota and Wisconsin will share data. So, it can snowball to an extent. But, in theory, if all the states had gone to them, they would've had... You can have payment plans and things like this, but ultimately they would've owed $25 million and they had a million dollars in the bank account, so it could bankrupt your business if the liability is big enough.
Turner Novak:
You can technically just continue to not pay, but they'll add more fees and then eventually will they like, "You can't operate in the state anymore," or something if you don't pay.
Sam Ross:
Yeah. If the state what reaches out to you, they can put a lien on your business.
Turner Novak:
What would happen there? When business that operates in the state, they have the first right of refusal to the cash that gets collected or something like that?
Sam Ross:
Yeah, actually, to be honest, I'm not totally certain. I've seen people with bank accounts get frozen in some cases. But, if it's a lien from a state that you're not operating out of, I'm not exactly sure of the recourse. But, I've definitely seen people's bank accounts get frozen before, or they'll send scary letters to Stripe or Shopify, and certain things can actually... They might be forced to freeze your accounts.
Turner Novak:
Yeah. Okay. So, it's like one of those things where you do need to handle this and you do need to take care of it, because it will eventually stop you from being able to operate your business.
Sam Ross:
Yeah. And especially if you're ever going to do any audit on your business, that's often where this comes up as well. So, I had to call someone, they're doing an M&A transaction and there were some historical exposure, and what happened was they hadn't... They weren't a client, they were just trying to become a client of us last minute, and they hadn't been collecting in certain states. And so, the auditors are saying, "Hey, it looks like you have liabilities of $2 million in different states." You need to put $2 million under the sale price into this collateral account and market it off the deal price, because that's a liability that I don't want to accrue as I'm buying your business. Now in some cases you're doing an asset sale, there's ways to smooth some of this over, but in this case it was a stock sale and so there was no way around it.
Turner Novak:
I was going to say, I've seen that before. I've had portfolio companies that at the end of... They sign a term sheet, they go through a two-month diligence process, and then the second to last day they're like, "By the way, we're knocking the price down by your accrued liability in this one place, because we think there's outstanding sales tax exposure."
Sam Ross:
Exactly. Any buyer business, they always... They have a checklist of things they can do to knock your price down. And so, sales tax is just one of those things on the checklist and it's a pretty easy one to just find something there. And so-
Turner Novak:
Because, customers are usually just willing to pay, it's just an extra two, three, four, five, 6% just at the time of sale that you just kind of bolt on and they usually pay it.
Sam Ross:
They have to... Yeah, they have to pay. I think the nice thing is in the US at least you don't show it until you're at checkout. And so, it's hard to know. You can't really run an A/B test around this to know how it affects your conversion rate. But, I think, customers are used to paying sales tax, and so, yeah, you just have to do it. The thing is it's legally mandated, right?
Turner Novak:
Yeah. So, one thing too, I think you guys have been really focused on building out good infrastructure, and this is a podcast, it's a tech podcast, we have to talk about AI. I feel like we're contractually obligated at this point. How are you using AI at Numeral?
Sam Ross:
Yeah. We use it in a handful of ways. I think one is, again, a lot of these bigger companies, what you end up doing is you get to scale is... You have these thin tools, incumbent tools like Avalara, and in order to manage them, you either hire a bunch of staff, someone's like a controller, head of tax, or you hire a consultancy like a sales tax consultancy to manage that software, because there's a lot of complexity. And so, what we're trying to do is make it so that anybody in the company can manage sales tax without much expertise. And so, all this more service work goes away. So, to give an example of that, as I mentioned before, you get all this physical mail from the states, they'd love to send it to you every week, you're registered in 40 states, you're going to get some mail.
What we did early on was we said, "Let's scan all of that mail from our customers and we'll read all those letters and if some action needs to be taken, we will just go and do it for them so they don't have to think about it." So, you might be filing... A really common thing that happens is if sales tax are typically filing on a quarterly or monthly basis, but what will happen is maybe you're filing quarterly at the end of the year, the state does an analysis, "Hey, their sales have gone up, we actually want them to file monthly. We're just going to send them a piece of mail that says now you file... Now file your..." Michigan's going to say, "Hey, now you've got to file monthly in our state."
Turner Novak:
And to not only miss that.
Sam Ross:
Exactly. And so, Avalara's not going to update automatically. And so, if you don't... Usually you get the piece of mail you have to go within a few days of opening that letter. And this is a really small thing, but this mistake happens constantly. You forget to do it, and all of a sudden when you go get file your taxes at the end of the quarter, you now have a bunch of delinquent filings you didn't have with penalties, and you have to go... And there's just a bunch of work to get that fixed. And so, we scan all that mail automatically. We started off, we had some contractors just reading thousands of pieces of mail by hand and categorizing it, and it was a horrible job. They hated it.
They had this beautiful spreadsheet, like Google Sheets, but it was low brow. And honestly, it's nothing too fancy, but it was a handful of AI prompts and things. We're actually now able to automatically sort all that mail to different buckets, and we actually automatically take action. So now, yeah, if Michigan tells you you have to file monthly, there's no human intervention, we were able to read that mail, parse what it's saying, and update in our system that, "Okay, now client Y has a monthly filing in Michigan." Or if something that's escalated, we know to put it to an escalation team. So, there's still maybe humans at different points, but it is able to do all that reading on our behalf and take action.
Turner Novak:
So, it's not like if I'm a user of the product, I'm getting this generative AI LLM powered experience in the interface, but it's on the backend of just like we no longer have this contractor army of people who don't actually enjoy what they're doing and just we've automated it all and just shows up in the software.
Sam Ross:
Yeah, exactly. I think the user experience that I always aim for is that you just barely ever log in. And so, a lot of it is behind the scenes. The other thing that's pretty interesting as well is I talk about these different tax rates and all these different localities, and the question is, "How do you actually get those updated rates?" Making this example up, but how do I know that Juneau, Alaska, is passing a 7 cents tax on soda cans or something like that? And it's not like there's some centralized ledger database that all the states are updating with the soda can tax laws. It's like they'll put out city council notes, they put out the postal law somewhere, and the way that the incumbents have solved this is they'll have huge research teams. Like, Avalara has over 50 folks who are just manually going and reading laws, each month they just go back to the law and they just read through it, hitting control F, trying to see if anything changed there. And they make mistakes all the time.
And so, this is an obvious place where: What are LLMs good at? They're great at reading data, there's great at reading written text and extrapolating what it means. And so, a combination... We're using a lot of these RAG engineering techniques to go take all the laws, reference them, and then there's still human evals going on in our side, but able to build this, we call it tax research, that normally would cost tens of millions of dollars and require armies of lawyers and researchers, and doing it at a fraction of the time and a fraction of the cost with higher accuracy. And so, it's adjacent to maybe some of the techniques that a lot of these legal AI startups are doing.
Turner Novak:
Yeah. And it's interesting, it's one of those things where you're removing the human work. Someone could get really upset that you're replacing the jobs or whatever, but the people are control F-ing at PDF and find... Come on, no one enjoys that. We should be automating that job and move those really talented people that understand tax law into doing things that are maybe tax law, maybe, I don't know what, doing things are better for the economy but are better for the world. But, you're like upskilling... You're able to put these people into things that just are more enjoyable or add more value.
Sam Ross:
Totally right. So again, we have our head of tax research. He was a senior director at Avalara, and he was used to just doing a bunch of control F, and reading... And doing really... He had a team too doing this, but a lot of low level work. And honestly they were just moving those jobs to India anyway, because it was often lower value. But, there are cases where we really need someone like him. He's a trained lawyer to go look at this and is this... And again, is it edible underwear apparel, or is it candy? And someone who's a trained lawyer with years of experience is probably going to have... Is going to have a more nuanced take on that. And so, it's not that the high skill human parts are eliminated, but yeah, there's a lot less rote, boring, mind-numbing work going on, and the low level stuff has been automated.
Turner Novak:
So, if I'm somebody who I'm competing against incumbents with let's say a less good product. I don't know if this is a fair way to set this up. But, what advice would you have to them of here's how I would approach selling the product and converting people over?
Sam Ross:
Yeah. So, early on our product was not good. I didn't have a dashboard for months. I remember so many emails like, "Hey, we're refactoring our dashboard. It'll be ready in two, three weeks." And we were just doing all of our compliance... It was basically a service. Not every business can do this, but for us it was a service business. I'm like, "Send me your data over an email and I'll send you back your tax filing over email." But, the one thing we did do that the incumbents would, we're doing was just going above and beyond. You could book a calendar, you could book a call with me anytime, I respond to your emails right away, I was reading your... I was the one reading people's mail at the get go.
Turner Novak:
Did you sit in the customer support email just constantly answering?
Sam Ross:
Yeah. Constantly answering. I was doing all the sales calls too, so everyone knew me. And I still do this today. Every time I jump on a customer call, I give my personal phone number on every call, because I want to know if something's not going wrong. We have amazing customers experience team and it shouldn't get escalated, but if you have lien on your business or something happened, I'd much rather someone complain to me and I know and I can solve the issue than just someone's just frustrated and quietly mad or going to churn or something like that. So, I think, what you can do is it's the boring refrain of do things that don't scale. But, I think, the other thing too is I think in this age of AI, what we really figure out you can do is you can basically start by building to service business out of the get go.
And then what you do is you figure out how do I take these things that are more service looking things, like reading people's mail, and actually turn that into... How can I use AI to automate that? So, another example would be like you need to figure out what's the product category here. So again, I was saying you look at something like an Olipop, like those healthier sodas, and it's like the client may make... The way that incumbent software is they say, "Hey, go apply tax codes to your thing." So, you're a customer of Olipop, I go to Avalara, I go, "I guess I'll put soda there." But actually... I'll put the soda tax code 257. And so, now I know it'll be taxed as soda,
But, actually what we'll do is to say you shouldn't be... That onus should not be on you. You're not the tax expert. You don't actually know what the taxability is. And so, what we'll do is, by hand, again, we were doing this first we'll make sure, "Okay, Olipop, no, it's actually... The tax category is sparkling juice with zero to 25% juice is the taxability for that," right? Because again, based on different laws, some states tax it differently. So, first of all, we prevented that mistake from happening at the get-go. And we were doing that by hand, and we were just like... I was doing it because I didn't trust anyone in my team that... We trained some folks that have sales tax expertise. And then now we have multimodal LLMs. We actually will make sure if there's nutrition facts or supplement facts, let's get uploaded into the LLM. And it is not that it's 100% automated, there's hard cases where maybe someone has to come in and really dig into and do some maybe original tax research. But, a lot of the low-level stuff can be automated. But, it wasn't at the start.
Turner Novak:
It reminds me of one of my portfolio companies is called Hanover Park. They do fund admin. I don't know if you've ever invested in a venture fund or any of these private asset classes.
Sam Ross:
Yeah. Yeah, of course.
Turner Novak:
So, there's some software for it, but it's still a very manual thing. If you use Carta, you get on a sales call with Carta, there's five salespeople on the call.
Sam Ross:
Oh my God.
Turner Novak:
It's still pretty manual. And he's doing a similar thing with fund admin services, of trying to just automate some portfolio updates. Your portfolio sends you an update and it ingests it in and it's in the software of how the company's doing. It's one of those things where it's a little bit scary when he told me that this what he was doing, I was like, "I don't know if fund admin, everyone hates their fund admin provider." But, then it reflects back on like, "Maybe there's an opportunity to use LLMs to automate some of this manual low-level work that has historically given a bad product experience." But, you can actually flip it, and by automating some of the service aspect... It's still always there. Customer service is a big component. Like for him, when somebody... For a customer, he guarantees the implementation, he goes, and he'll sleep in your office until it's live. So, it's like they're automating the manual stuff, but he's still hardcore just like, "Whatever you need me to do, I will do for you."
Sam Ross:
Yeah, exactly. I think there was a wave a few years ago, I don't know if you remember where there was tech-enabled services, you had Pilot, and then Justin Kan had his AI law... Or not AI-
Turner Novak:
Atrium.
Sam Ross:
Atrium, the law firm that blew up.
Turner Novak:
Yep.
Sam Ross:
And I would say maybe it was the idea wasn't wrong, but maybe the timing was wrong. And, I think, there's probably still cases where it doesn't work. I'm sure. Again, our business, one of the things that's nice, sales taxes is typically most decisions are deterministic, there's not a lot of interpretation. And so, the stuff can be mostly automated. But, I'm sure there's a bunch of other services business where maybe you just start off by building it, like the VC... A lot of VCs are going to hate this, but you build it like a normal service business and then you just go workflow through workflow and find ways to automate it. And so, my guess is there's a whole host of unicorns, sitting ideas, that are there that waiting for that. And ideally you build the automation from day one, but you can always do things that don't scale at the start.
Turner Novak:
Yeah. And speaking about VCs, I think it's good to get their feedback on things and they may have opinions on... They've seen just tons of companies, and it's sometimes you can fast track and avoid certain dead ends. But, also they're not the all-knowing experts. They're wrong all the time too.
Sam Ross:
Yeah.
Turner Novak:
But on that note, so I know that you've raised a couple rounds for Numeral, how did you think about raising money? Because you mentioned before, bootstrapped, you like owning the business-
Sam Ross:
Yeah.
Turner Novak:
And not... You're working remote, nomadic, you like not having the stress. How did you think about, "Okay, I'm moving back to San Francisco, I'm going to raise money for this thing?"
Sam Ross:
Yeah, good question. I mean, definitely some of it was just being ignorant or something or lying to myself a bit.
Turner Novak:
Okay.
Sam Ross:
I would say I had, so again I was at my company, Teespring and that company was valued at whatever, half a billion dollars. Two of the founders are close friends. One of them was in my wedding and they were like, we're in early twenties, they're worth nine figures on paper. I'm like, "Holy crap." And then it just went through some traumatic stuff where everyone got wiped in the end. And so, I definitely have some scar tissue from doing these venture businesses and seeing the toll it can take on just your mental health as a founder. And so I think, I'd always kind of like, I knew I wanted to do it eventually, but kind of kept kicking the can. And so we started looking at this business, Numeral and we're like, "We could probably get it off the ground bootstrapped."
But ultimately, I think for some business models like SaaS, the venture model works really well for a reason. You have all these big upfront fixed costs, like the engineering and on the sales side, because it's a highly retentive business, it pencils out. So we're like, if we just do this the bootstrapped way, we're just going to go way slower. Our product's not going to be as good because we won't be able to invest in R&D and engineering as much." And if it works, then someone's smart's going to just go raise a bunch of money and that'll compete us there. And so they're like, "What...?" We're kind of somewhat forcing that decision. And I think we also have always tried to take money incrementally. So don't take much more than you need and you can always go back, if you really need more money, you can always go back out and get more.
And so, that was my approach. I'd always wanted to do YC. And so it was kind of an easy way to just, you start the heroin drip of VC funding there and it feels all safe and cushy because YC is a community and then before you know it, you're just on that treadmill. But again, I think if you're going to build a serious SaaS business, I think it's really hard to do it. So obviously, there's cases where people have done it without, but founders, I think startups are getting more and more competitive and there's more and more venture money out there. And so someone sees that you're doing a good job. We've had people come and copy our business model. Yeah, you just have to be like, capital is a way to be just more competitive.
Turner Novak:
Yeah. I know you work with Susan at Uncork. How did that come about, meeting them and meeting her?
Sam Ross:
Yeah. After YC of the demo day or even before demo day, you have all these investors come through and she had a thesis around the space already. And so yeah, we meet a lot of investors, but I think she had just high conviction, done way more research on the space, had talked to the founders of someone in our space like TaxJar and just really understood our vision well and so resonated. And then I think also, we decided at that time too, it was our seed round. We thought it'd be better to have a seed specialist versus someone like a multi-stage firm, which can create some risk if they don't want to invest you with the A and maybe have less optionalities with the A, having a multi-stage firm do your seed. And so we thought at the margin we just, I think liked her on a personal level and seemed like she would hustle for us, and so that's why we decided to work with her and she's been great.
Turner Novak:
Can you explain that a little bit more? What's the risk of a Series A firm or multi-stage firm doing a seed, not doing Series A? How does that play out or if I've never heard that before, why is that important?
Sam Ross:
Yeah, so I think it's... Effectively what happens is, these multi-stage funds, they're doing seed checks. They're basically putting an option on your business. It's like the option to do the A later on. And first of all, when you have a huge fund, hundreds of millions of dollars giving someone a 2 million dollar check, they don't really care about you.
Turner Novak:
It's like 0.1% of the portfolio. They need to do a thousand of those to deploy the capital.
Sam Ross:
Exactly. So they'll just give you a check. They don't really care about you.
Turner Novak:
Yeah.
Sam Ross:
And then second of all, when you go to Series A, there's kind of an expectation that they'll either lead the round or minimum participate. And so, if they don't want, for whatever reason, they don't like your business or there's some disagreement there and they don't want to lead or do your A, then it can make it very difficult to go and do your A because it's a huge negative signal for the business.
Turner Novak:
Yeah.
Sam Ross:
And also, you might be more constrained too, in terms of valuations and things like that because they have sort of this unfair leverage on you.
Turner Novak:
And so the reason you do somebody who specializes at seed is because it's, like from someone like Susan and Uncork, it was a meaningful check for them. If you do well, it moves the needle-
Sam Ross:
Yes.
Turner Novak:
... for them and she'd be more willing to jump on a call to do something to help you go.
Sam Ross:
Yeah, she's a vested interest. And then I think even at the A, she's probably more unbiased person in terms of making introductions and things like that because she's not going to, maybe they'll do their pro-rata, but they're not going to put in a sizable check and so they might be a little more objective in terms of who you get at the A.
Turner Novak:
Yeah, makes sense. I think when you raise your Series A, I think you said it, you did it in a week or something?
Sam Ross:
Yeah.
Turner Novak:
What's the story with the Series A? How did you get ready for it and then how did it all come about?
Sam Ross:
It was funny. So yeah, there was a little bit of preparation. You want to get your financials right and I actually, I'd gotten married right when we started the business and so I couldn't take a honeymoon because we were like seed stage. And so a year later, I had a deferred honeymoon. I was in Japan, but we did have a in-house on finance team and our financials were in messy states. I was waking up every morning stressing about making sure our ARR was clean and getting the financials in a good place.
Turner Novak:
On your honeymoon?
Sam Ross:
Yeah, we were like waking... We had the nicest hotel of my life in Hokkaido. It's thousands of dollars a night and I'm still dreaming about my ARR and how do you think about committed ARR versus pending ARR. ARR is, it's not a GAAP compliant metrics, so you can think about different ways and just stressing me out, trying to take calls from there. And so I kind of came in pretty, a little bit... Business was doing quite well, but there was just some stress around how buttoned up are we?
Turner Novak:
Because you don't want to lie about that stuff. You just want to make sure that you're clear about what all these things are.
Sam Ross:
Yeah.
Turner Novak:
By the way, it's like my wife's nightmare of us on a vacation and me just thinking about work.
Sam Ross:
Yeah, we literally had three had baths in our hotel room. It was a super premium thing, the nice thing ever. And then in the bath, my wife's getting mad because you can tell my brain's thinking about-
Turner Novak:
Yeah.
Sam Ross:
Or I might get my laptop. That's founder life, I guess. But we came back, we got into a good enough place to go out and I went on... On a Tuesday and I met Chetan from Benchmark and within, I think, 10 minutes of talking to him, he understood everything instantly. And he had seen the space before, he'd been to VC for almost 20 years and had looked at Avalara early in his career when it was, before it went public. This is probably in 2008 or 2009 or something like that, and always thought the space was underserved and understood the AI angle and all of that.
So 15 minutes in, I got a real lesson in terms of selling from him. He's like, "Wait, wait, wait, pause, pause." I mean, again, he's from Benchmark, he can do this. He went and he's like, "Open your calendar." I'm like, "Okay." He's like, "Let's get dinner tomorrow night." It's like, "Okay." "And then block off this time on Friday. I mean, we should do a partner meeting on Friday." I'm like, "Okay." And then he's like, "Okay. And then also, what's your phone number? I'm going to send you six reference checks on me right away."
Turner Novak:
Oh, wow.
Sam Ross:
Went full, a full court press and it was just really aggressive selling. And again, being someone from a highly reputed firm, things like that, you can maybe be a little more aggressive. But it was definitely a learning lesson for me now, when I even go and try to close really... Candidates that I'm really excited about, I've had to learn to just go... I fly out and get drinks with someone and really just do this, give reference checks to myself, do this kind of full court press. And yeah, I think, again, there was kind of alignment from day one in terms of what we were trying to build and we were fortunate to have good team, good metrics, all that. So yeah, we basically did our Series A in four days and ended up having to go-
Turner Novak:
And you met Chetan on the first day?
Sam Ross:
I met him on the first day. I guess you got intro. I forgot to put him, I forgot to get an intro to him. And then I actually, Susan, who's on another board with him, intro'd me that morning. So I got intro'd to him that morning, took a call in the evening and then-
Turner Novak:
Dinner the next day.
Sam Ross:
... dinner the next day. And then on Friday, I was staying at my... Because I was living in LA at the time and my parents live in the East Bay, I was staying at my parents. They saw me, I've been working until 1:00 AM every night. They saw me exhausted like, "Oh, aren't you excited to have the weekend here? It's like Friday night. You better relax." And then I'm like, "Actually," I was texting with Chetan. I'm like, "I'm actually going back to the Benchmark office at 9:30 tonight on Friday night because we were going to go and just negotiate this thing out right now."
Turner Novak:
Yeah.
Sam Ross:
And so yeah, by the time we... We just sat in a room and worked through the numbers, I think 11:00. We're like, "All right, good enough." We agreed on terms. But it was nice because I really wanted to get back to work too, and I think I could have gone longer. We had a lot of interest. You could go try to maximize your price and drag these things out. But I think as an entrepreneur, you have to just be decisive and move fast and not... I see a lot of folks that just like if you're highly neurotic, you're not going to be a good founder because you just need to just commit to decisions.
And so for me, I think there was mutual respect there that we were both... He as well, I had some VCs that were going to my data room, which again was like, I wouldn't call it a data room, just going to my spreadsheets, which were, the data was fine, but it wasn't as... This is not public company numbers and they were obsessing over random details and I love that Chetan, he asked me some questions, but he wasn't focused on, are we going to go from a Series-
Turner Novak:
Are you going to be at 7 million or 8 million ARR by the end of the year?
Sam Ross:
Yeah. I said, "Are you going to double?" Some folks are like, are really obsessed like, "How are you going to go from current ARR to 3X the size?" He was just much more concerned, "Can you actually build a 10 billion business doing this and do we have alignment on how we're going to approach that?" And so, it's really seeing the forest and the trees. I think it's important. All these other firms that'll bring in some associate who's fresh out of banking and all, they're breaking down your P&L details. And I think maybe for a later stage investment, that makes sense. But for a Series A, that's not how I would approach it, at least.
Turner Novak:
Yeah, interesting sales lessons. To your point, when I'm on a call with founder and I know, "Oh, this is pretty interesting, I'm going to be in New York next week, let's get coffee when I'm in New York," or whatever the extent of when the next touch point in the sales process has to be.
Sam Ross:
Yeah.
Turner Novak:
Yeah. Definitely something I've learned over the past. My side fundraising with LPs, but also with founders too. I was just like, "Okay, I know this is interesting to me. Let's just move forward and show that there's intention there," instead of just kicking the can down the road and doing stupid shit that doesn't really matter at the end of the day of that nitpicking about certain data room points or whatever. I think there's this quote of, every deal dies in the data room or something like that.
Sam Ross:
Yeah. Look, I mean, there's fraud out there. And so you want to make sure, obviously, you're catching that. You're not investing in those things, but I think yeah, you have to be decisive otherwise because the end of the day, everything's about speed, right?
Turner Novak:
Yeah.
Sam Ross:
And so this is why I like the big companies. Someone like Google is like, someone behind OpenAI, despite probably having a lot of real structural advantages is because you have a team of eight people being neurotic about covering their ass and all this other stuff, where it's like Sam can just come in the meeting and be like, "We're doing this. I don't give a fuck. Let's just go. There's some risk, but I don't care." You just move way faster. So I think it's, everything, just making decisions, if you don't have that level of confidence to just... And willingness to... Also, sometimes you're going to make mistakes sometimes and you're being just and quick and decisive, but as long... You have to just go quickly.
Turner Novak:
Yeah. This is actually a question from Nate on your team. So I mean, it seems like Numeral's firing on all cylinders, working really well. If you had to look forward five years and you guys fail, doesn't work out.
Sam Ross:
Yeah.
Turner Novak:
What do you think the reason would be?
Sam Ross:
Good question. I think a couple of things. I think one is broadly the way I see it today at least, is that everything's really an execution question for us. The market, I think, is there for us. There's been multiple public companies worth roughly 10 billion dollars here. This is not some novel space we're trying to create-
Turner Novak:
Every company that generates revenue has sales tax, pretty much.
Sam Ross:
It's legally required. And then the bigger vision again is, nobody's built a global version of this. So we look at deal all the time despite maybe what's going on in the news. What they did figure out was when you had payroll before a deal, maybe you use Gusto or Rippling or whoever for US, and then you're stuck using a bunch of different random global vendors, I need my local accounting firm in Japan to handle my Japanese contractors. I need this nasty bundle of vendors and they're just like, "F it. We're just going to go to these countries, do the hard work of creating a more vertically integrated business and getting lawyers on the ground." And that's actually what we're building as well is actually going and so it's kind of a one-stop shop to get your global compliance.
And so I think this is a really hard thing to do though, actually going and figuring out how to deal with the Korean tax authorities is not simple. In Mexico, you actually have to go often and to file things, you have to send someone to this building in Mexico City, they wait in line for an hour, you need a wet signature. There's a lot of just nasty operational stuff. So I think just failing to execute is always the thing that least scares me the most. It's nothing structural. I think it's when things are outside your control, it's less scary because you can be a little bit more stoic about it. "Hey, I can only control what I can control and everything else, you have to just be a little zen." But when you know it's in your control, then that's what keeps you up at night. Because yeah, I think that that's probably the biggest thing that would probably be the reason we fail.
Turner Novak:
How big do you think Numeral could become?
Sam Ross:
Our ambition's to be a decacorn at some point, a public company. Again, I think, you look at these businesses like Avalara. They're probably purchased by private equity, but they got taken private at 8.5 Billion and they're 90% of revenues for US tax. And they'll put out things, the TAM is 30 billion of this market. So I think certainly, it's become worth at least 10 billion dollars. And I mean, that's maybe a reductive way to put it. I think ultimately, again, we just want to build the first piece of software where you have one place that handles your taxes everywhere and just kind of makes this whole thing of sales tax or the broader term's indirect tax, kind of melt away. I have a bigger galaxy brain vision for it, but I don't know if that's a-
Turner Novak:
Yeah, what's the galaxy brain vision?
Sam Ross:
So my galaxy brain vision of it is, you talk to a lot of economists and what they'll tell you is, actually income taxes are pretty bad because income tax, what it does is it disincentivizes savings, disincentivizes investment. When I say income tax, I'm talking about capital gains tax, corporate income tax, personal income tax. So you're really taxing people for income, for capital gains, you're taxing people investing. And really, in some ways what you would actually rather be taxing is consumption because that money you're using to consume could have actually been used as an investment.
Turner Novak:
And there's a lot of countries that mostly have consumption taxes, right? Instead of income taxes.
Sam Ross:
Like Singapore, et cetera, right? The challenge with higher consumption tax, and again, consumption taxes, sales tax, things like that, is that it tends to be regressive, meaning Jeff Bezos and Elon Musk and myself are all going to pay this same amount of tax on our coffee from Equator right now. And so, it ends up probably under taxing. If you want a tax system that's more progressive and taxing wealthy people, it's going to under tax them. And there's been a bunch of papers over the years in terms of how you could design a more progressive consumption tax. The challenge is, it's a pain in the butt to implement. So how do I have a special tax rate so that when Bezos goes and buys his 2 billion dollar yacht with his wife on it, how do we know that... How do we tax that at 70%, not the same 6.5% that your soda can is taxed at.
It just makes the tax code so complex by being able to tax certain goods differently. And we have some of that, but it's tractable to an extent. And so my hope is, you having software like us empowered by AI, makes the cost of compliance go down so much and allows us to actually add a lot more complexity to the tax code, which most people don't want to hear that you're adding complexity. But I think the hope with AI is you can actually, the complexity doesn't become as much of a cost anymore. And so then now, you can actually design these consumption taxes to be much more progressive. And then over the long term, the real goal is not to raise taxes, it's just to get rid of income taxes, get rid of, and so it actually allowed the economy to grow much faster and you kind of have a system that has much better incentives in there.
Turner Novak:
That's definitely the galaxy brain vision.
Sam Ross:
Yeah. Yeah, yeah. I'm not putting it in my, if I pitched an investor and I think I don't want to bring that in and go on some 10-hour libertarian rant or things, but I think about it when I'm in the shower late at night. Something I do think about.
Turner Novak:
I mean, what I was thinking about a little bit was if I were to say, "How does this get a much bigger TAM?" The thing I think a lot about is how could this TAM actually be way bigger than what people think? So there's the Salesforce example of I had Villi Iltchev, who was the head of Corp Dev at Salesforce for a while, and he's like, when he first joined, the market size of CRM software was like 8 billion or something, and now, Salesforce does 30 billion in revenue or 40 billion.
Sam Ross:
Yeah, that's crazy.
Turner Novak:
10% market share. It's like the TAM was way bigger. So I guess for something like sales tax, you could think, okay, maybe e-commerce is only X percent.
Sam Ross:
Sure.
Turner Novak:
They're using that to value the sales tax.
Sam Ross:
Sure, sure, sure.
Turner Novak:
And then on the software side with AI, does more labor shift to software so the sales tax actually get way bigger?
Sam Ross:
Yeah. So I think there, right, you look at the economy, services is usually the biggest chunk of the economy. More and more of that's probably going to, I would imagine it moves over to-
Turner Novak:
To software.
Sam Ross:
... software, AI. And by the way, software is only... Software is tax-exempt in California. It's only taxable... It used to be taxed in two or three states. Now, or started at zero, obviously when software didn't exist. Now it's taxable in roughly 30 states. And you imagine over time, as software becomes a bigger percentage of the economy, it's just going to be taxed everywhere. And the other thing is, businesses are going, despite maybe some of these short-term macro trends, businesses are going global right away, right?
Turner Novak:
Yep.
Sam Ross:
So if you're an early stage AI startup company, all these companies are claiming these crazy ARR growth.
Turner Novak:
Yes.
Sam Ross:
They're selling in Azerbaijan and Albania and all sorts of weird places. And when we talk to them, they're like, "Yeah, I actually have all these liabilities in all these long tail esoteric countries and I don't want to think about that." And so, as businesses are A selling globally right away, AI driven software become, hopefully eats into services economy and just grows the pie, then obviously we'll grow linearly with that or it'll correlate with that.
Turner Novak:
Yeah, it's pretty crazy to think about. So as it all gets more complicated, you want to make it simpler for people at the end of the day.
Sam Ross:
Exactly. I think, again, the promise is, we do a lot of nasty stuff underneath the hood and the nice clean stuff, maybe AI automates, the gnarly stuff as you send someone to Mexico to wait in line to do this thing.
Turner Novak:
You're still doing that?
Sam Ross:
Yeah, yeah, yeah. There's no, I mean, you can't AI away a government law. I mean, maybe advocate in Mexico one day to get rid of wet signatures, but there's a whole host of stuff we do underneath the hood and I think, but the goal is you never have to think about this stuff. It's all obfuscated away and we're doing global money movements as well. And so, I think all these things as well, at least from my selfish business perspective, you find a ways to charge for all that as well.
Turner Novak:
Yeah. So then a little bit of a more calm, simple, easy question to end it on. You've lived nomadically for so long, now you're in San Francisco. What do you think of San Francisco?
Sam Ross:
Ooh, good question. I also grew up here, so I have some weird takes. I would say, let's see, know, I think once San Francisco was, when I first moved back after college was like, it was 2011, 2012. SF had this amazing energy. You had the Mission was full of service workers and artists and people. You go out on a Tuesday night and there's all sorts of bars are busy, and now it's much, I would say, a lot of that's left. It's a lot sleepier. I think people complain about all the dirtiness, the homelessness. For me, that doesn't actually, I mean, I want people to be taken care of and stuff, but that's not what actually bothers me. What bothers me more is that it's just the boring, it's gotten much more boring and lack of cultural things going on.
SF's a good town for business. And I think, again, we were talking earlier, the Tenderloin, places like that, it's actually cleaned up and stuff, but I actually don't know if that actually solved... It doesn't solve the problem that I remember in 2011, 2012, people moving from New York to San Francisco because it was actually on the same footing culturally in some ways. And all these hot restaurants were opening and there'd be good shows. Just because the city gets cleaner, it's still just going to be a bunch of techies like me who are drinking non-alcoholic beers on a Tuesday and going to bed on time.
And so I don't know how it gets that... I would hope there's ways to get cheaper housing and things like this to fix it. And I think you have folks doing the YIMBY stuff, which maybe helps with the margin, but it's just really hard for me to see how you can bring housing prices down enough to get back people who are really contributing more culturally to the city. Otherwise, you end up somewhere, it's more like a Singapore or something where it just feels like these kind of transient business cities.
Turner Novak:
It doesn't have a soul almost. You want to bring the soul back to San Francisco.
Sam Ross:
SF's whole thing was it had a deep soul, right? In the sixties, you had all these, it was this blue collar city. Then in the sixties, you had all these hippies move here and they made it a weird city. They made it okay to be gay. That's really, I think, what brought a lot of the specialness of it here was just this very open-minded place and very accepting, that maybe brings some things that are less pleasant, maybe more openness and empathy to homeless where things get dirty or things like that. I think now, it's kind of weird going from the place that had the very distinct culture to feeling more like I could be anywhere in the world right now. And it's the same generic shops. And so I don't know how you fix that, but...
Turner Novak:
Because I feel like that energy, the willingness to try new things, openness to new ideas, is really what it takes for a successful startup ecosystem. So you could maybe... Maybe I'm being critical here, but if you click and drag out, you lose the soul. You lose some of that creativity and that the bold willingness to try new things that makes San Francisco what it is for tech.
Sam Ross:
Yeah. Before I started this business, I was living in LA. I still kept my place there. I split time. And what I love about... LA is ugly, the air quality sucks and nature's much worse, at least where I live on Silver Lake on the East side. But what I love is, there's always way more going on. I went shopping the other day, there's this weird pop-up, it's like East side LA tennis club and there's no business model here. They're just selling weird tennis club stuff that's people that don't actually play tennis. And I'm like, this pop-up could not survive. The rent would be too high in San Francisco. People would be weirded out by this weird clothing.
But I love it. It was so funky. Or my wife and I love going to Koreatown and eating in a hole in the wall restaurants. I don't think you have... You still have immigrants coming from Korea, immigrants coming from Central America and opening restaurants and adding to, Armenia, adding to the culture. In San Francisco, the only immigrants coming are engineer tech workers who, again, it's great that you can spend money in the city and contribute stuff, but they're not, I would say... Someone like me, I don't add culturally to the city too much. And so, I feel like I'm part of the problem for sure.
Turner Novak:
Yeah. You think of when someone moves to San Francisco, where do they move? They move to Hayes Valley, the center of all the AI stuff or I guess the Mission now, right? Move next to the OpenAI office.
Sam Ross:
Yeah, exactly. So I'm hopeful for, I think Mayor Lurie is doing some good... He's doing some good stuff, like public safety matters. But again, for me, what matters is like, is it interesting? Is it fun? Is there some cool new cultural thing I can check out? And I think the cultural piece, there's the money to solve it, but most tech people, if you get money in finance in New York, you might be going to spend a bunch of money on art and you're trying to buy cultural capital. The way to get culture in LA, same thing. The way to buy cultural capital in San Francisco is to do a bunch of angel investments. So maybe it's good for the GDP of the Bay Area, but it makes it a little less interesting.
Turner Novak:
Yeah, we need to make it cool to invest in culture in SF.
Sam Ross:
Yeah, exactly. Like burners. Burners are great, but not enough money going into arts. We do art. My brother-in-law is a successful ceramicist in LA and we're going to art openings all the time and that stuff's super... For me, I love going to environments where I'm the only techie there. And in SF, I go out in the night and everyone's talking about whatever-
Turner Novak:
The new model from... Yeah.
Sam Ross:
Yeah, Cursor's latest round and how did they do that? How are the economics of that? And that's fun to an extent, but I think it just becomes a little monotonous. And maybe I'm just self-conscious that I look and feel like everyone else here.
Turner Novak:
I think this is a good way to end it. This is a lot of fun.
Sam Ross:
Great. Awesome. It was so fun to catch up and thanks for having me.
Stream the full episode on Spotify, or YouTube, Spotify, or Apple
Find transcripts of all other episodes here.