Pod Insights Summation with Auren Hoffman Clay COO Varun Anand - why cutting-edge GTM teams are moving away from traditional CRMs About this episode Varun Anand is the co-founder and Head of Operations at Clay, a GTM platform that helps over 10,000 companies with data enrichment and sales outreach. Clay recently raised a $100M Series C funding at a $3.1B valuation. In this episode of World of DaaS, Varun and Auren discuss: The shift from seat-based to usage-based pricingWhy LinkedIn's data remains underutilizedHow AI agents are replacing traditional SDR workBuilding contributory data networks at scaleLooking for more tech, data and ventur...
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:02 Auren Hoffman? Hello, fellow data nerds. Welcome to World of DAS. I'm your host, Orin Hoffman, CEO of Incubate and GP of Flex Capital. Discover more episodes, get weekly data as a service news, original content, articles on data and more at worldofdas.com. That's world of das, D-A-A-S, dot com. Hello, fellow data nerds. My guest today is Varun Anand. Varun is the co-founder of Clay, a GTM platform that helps over 5,000 companies with data enrichment and sales outreach.
0:34 Speaker 01 We're at 10,000 plus.
0:36 Auren Hoffman? 10,000? Holy macro. Okay, I'm completely out of date.
0:41 Speaker 01 10,000. OpenAI, Google, Canva. Let's go. It's got to control.
0:45 Auren Hoffman? Whoa, okay, I love it. I know that you guys recently raised $100 million, Citi Financing, valued at over 3.1 billion valuation. Varun, welcome to World of DAS with your 10,000 customers.
0:57 Speaker 01 Thank you. I'm excited to be here.
0:59 Auren Hoffman? I'm excited as well. A lot of GTMs, they're coming together, dozens of tools to run all their campaigns and they've got maybe all these apps in the backend that are moving data around and stuff. How do you expect that to change over time?
1:14 Speaker 01 I think it'll change quite a bit. It might be helpful to just zoom out for a second on what the market is like, because I think it's an interesting market where you have Salesforce and HubSpot taking up a lot of the value with systems of record and massive companies. But then you don't have other real major businesses yet. There's a lot of point solutions and go to market. That's contributing to what you're saying, which is everyone has something for everything. Why is that? I think a big reason is a lot of these companies are founded by salespeople who found something that really worked for them and then productize it, which makes sense. The thing is, in go to market, you need to keep changing, because if you find something that works and other people copy it, then you have to change again. And then when you change, you can't use that product anymore. You have to do something else. What we think the market actually needs
1:58 Speaker 01 is this place we call like an IDE, where you can come up with new ideas, iterate really quickly, experiment really quickly and execute them really quickly. We think that's Clay, that place for people to do that. But we also think that's what the market is pushing for people to actually need. The problem of stitching together all these things, stitching together all of these data sources, that can be solved with something like that. That's kind of where we think the market is going.
2:20 Auren Hoffman? There's still going to be all these great point solutions to do something really amazing. Again, it might not only work for nine months. And then the idea is that you have this idea and it kind of hooks into this other solution. Or do you think the idea itself is going to be creating each of these things every few months?
2:39 Speaker 01 I don't think that the point solutions will last. You can see companies like Champify and User Gems and other point solutions like that, which are just telling you, hey, someone changes their jobs or someone came to your website. I don't think point solutions like that are going to endure in this transformation. And I think they'll become features of some larger platform, maybe Clay, whatever. So I don't think point solutions like that will endure. But what I do think will endure are individual data providers. There's going to be lots of niche data points that someone's going to spend the time on to go find and get. There are a lot of things like that that will kind of endure, but I don't think the tactical point solutions that we were just talking about will endure in the long run.
3:15 Auren Hoffman? A lot of these data providers are built on LinkedIn, some sort of way of crawling it. How does that work? Why hasn't LinkedIn helped with some of these solutions?
3:26 Speaker 01 Well, I think they're trying.
3:27 Auren Hoffman? And terribly, their products are just not really very good.
3:32 Speaker 01 They have an extraordinary trove of data that they're sitting on top of. They probably by their own admission would tell you they have not taken the best advantage of that from product perspective. In that absence, hundreds of thousands of companies have popped up to try to do something in that space. And LinkedIn has been very litigious with that. And that's their right. They have these Terms of Service and they're saying that certain companies are not compliant with that. My understanding is there's been some rulings on this topic in federal courts. That says that basically if you're scraping public data, that's okay. That's allowed. Now, maybe LinkedIn probably still has an issue with it, but that's what the legal standing is.
4:08 Auren Hoffman? If it's indexed by Google, it's okay. But we're also moving to a world where everybody has an agent. Right now, if I want my assistant or my BDR to log into LinkedIn and do some research, that is allowed by LinkedIn's Terms of Service. But if I have my agent on my behalf, login as me and do some things as me, which we're going to be doing for everything, I'm going to be doing that to make my travel reservation. Why can't I do that on LinkedIn? But then LinkedIn says, no, that's not allowed on your Terms of Service. It has to actually be a human, which is crazy.
4:42 Speaker 01 It is confusing and I think that they are trying to figure out how to make sense of all this in this new world.
4:47 Auren Hoffman? Once you have the browser AI, which is definitely on its way, I'm already using Comet, which is part of perplexity. All these things are definitely on its way. I don't even know how LinkedIn even doesn't know it's me. It's an agent. Maybe it's me. Maybe I'm doing it with an agent.
5:02 Speaker 01 It's not clear. It's a messy area and we're watching it closely, obviously.
5:07 Auren Hoffman? But ultimately, they're gonna lose those things. Your agent is you. So you have to allow your agent to be able to go in and do it. Again, it has to be respectful. So they could put a limit on your searches. So they could say, Varun, you're only allowed to search 50 times per day at this tier. And if you want to search more, you have to get a higher tier. And that's fine. They can't say, if you have an assistant on your behalf, you can't search for you. That's kind of insane.
5:33 Speaker 01 I think that's reasonable.
5:35 Auren Hoffman? What these sorts of do you wish we have that we don't yet have, or that are coming, or that you're excited about?
5:40 Speaker 01 One area that's pretty under-resourced right now is basically, there's still a staggering number of businesses that are not really online yet. Things that are online have really good capture. But anything offline, local businesses, small businesses outside of the US, still doesn't have remotely the kind of data that I think that we're used to in our working with like tech companies. I think there's a huge amount of potential there. I think we're almost taking it for granted still. When we see Facebook's numbers of how many billions of people are online, maybe individually, but the businesses themselves haven't caught up to the standard of consumer yet.
6:14 Auren Hoffman? Obviously, if you're in the US and your business, you kind of need to advertise somehow. You need to be found. If you're a car mechanic shop, you want to be found on Google or Apple Maps. You have some sort of presence somewhere online to be found, but that's not true everywhere.
6:31 Speaker 01 Even in the US, I don't think it's true. We still take it for granted. That's why all these HBS grads are starting these search funds to try to roll up car washes and car mechanics and bring them online. I don't know if that would be happening if that was really up to par.
6:44 Auren Hoffman? What other types of data do you think we need or is coming?
6:48 Speaker 01 It just depends on the audience of who needs the data. There are still a lot of closed ecosystems that are hard to parse. Meta is a fairly closed ecosystem. It's hard to get data from there.
7:00 Auren Hoffman? It's still remarkably hard to like create a custom audience and advertise against a small number of people. Let's say you have 5,000 people you really want to advertise to. It's remarkably difficult.
7:12 Speaker 01 Well, we're working on that. That's what we're trying to do. We're trying to make it very easy to get to develop like a custom audience that you can do anything with. Maybe that's advertising, maybe that's sending them personalized cookie jars or something in between. That's what we're working on.
7:25 Auren Hoffman? We all know LinkedIn is super slow. It's hard to use, takes forever. The company that's even worse than LinkedIn is Salesforce. The UI is absolutely horrible. It's literally terrible. It's embarrassing.
7:38 Speaker 01 I feel like this podcast is just trying to get me in trouble with all the companies we work with.
7:43 Auren Hoffman? I can't believe all the engineers haven't been fired at that company. It's just so terrible. Why do we all still use it?
7:52 Speaker 01 Salesforce and HubSpot have these entrenched ecosystems and these entrenched modes. And have captured so much value because they're becoming these systems of record. But I do think that that's shifting a little bit.
8:04 Auren Hoffman? I use HubSpot and Salesforce. From like a UI user perspective, HubSpot is 30 times easier to use than Salesforce.
8:14 Speaker 01 But I guess my point is I'm seeing this change. Basically, I'm seeing the most cutting edge teams not rely on those CRMs. I see them rely on Snowflake or Databricks, have it as like a flexible store of data. And then maybe use a claim, maybe use something else to like orchestrate it. And then the data just finds the rep at the moment that they need it. Maybe they get a text message before they meet someone. Maybe they get a Notion doc with a briefing before they meet a prospect, whatever. And maybe they get a ping in Slack. I think we're moving to a world where people are using UIs less as reps and the data is just coming to where you're doing your work.
8:47 Auren Hoffman? Which makes sense because no one likes to go in and enter stuff. They want it just to be there. And maybe they want like a quick look up. What happened with this customer? Did we talk to them before? But maybe that's to just get pushed to your inbox. Maybe you don't have to go and find it somewhere.
9:04 Speaker 01 That makes sense. We've built automations for our own sales team that do exactly that. Let's say this was a prospect meeting and I was telling you after this meeting, I would get a Slack message or an email. That says, hey, great meeting with Orin. Here's like a follow up that you can send. Do you want to edit it? Do you want to change it? And then like automatically get sent, things like that. I'll wake up tomorrow morning and get a briefing of all the signals that happened the night before. Hey, who went to the website? And then do you want to act on it? Do you want to do this? Do you want to do that? But it's not coming in a new platform. It's just operating where I already am, like in Slack or Gmail.
9:33 Auren Hoffman? I love that. We've seen with a lot of our companies that the number of people in marketing broadly, and let's include SDRs in that. The number of people in marketing is going down. In many cases, we're seeing companies where they used to have 40 people in marketing and now they have 10 or something. And they're using vendors, they're using AI, they're using tools, they're using APIs, they're coordinating, they're not spending less money. Are you seeing something similar?
10:05 Speaker 01 I think we're seeing like a shift in how the structure has changed. Because I think if you look at a traditional sales org, you have SDRs who are prospecting, you have AEs who are closing deals, you have rev ups who are managing systems, you have marketing who's trying to create pipeline. They basically operate like an assembly line. But the issue with assembly lines is they're standardized and they're not built for sharing learnings and insights. And if you're not doing that, you're not going to win. We've invented this role called GTM engineer where they're building revenue systems using AI and automation. Basically, we see people like those in the center of this, using tools like Clay or whatever to build systems. They're doing the work on behalf of all the reps, all the manual research, all the manual work that's happening with reps. That gets centralized to one person, two person, three people. And they're doing all this work on behalf of everyone else and building systems that can scale their creative ideas.
10:50 Auren Hoffman? Five years ago, I would say 100% of good marketers were Zapier masters. And today, maybe they're using like a Pipe Dream or an Aiden or one of those types of things to kind of put these tools together. These are at least relatively technical people have transformed. In some ways, marketing is more of a science than engineering. Are you seeing something similar?
11:13 Speaker 01 I agree with you that marketers are becoming way more technical, but I might disagree on the science versus engineering point. We actually do see the engineering discipline being brought to marketing on the growth side. Because I think it's more the systematic mindset, the systems mindset, and how you can build a repeatable system that can keep driving pipeline. And maybe you can even call that science because it is in some way. We see that systems building approach and mindset way more.
11:37 Auren Hoffman? When the salesperson's on a call with a client, we're already starting to see AI as a helper. You may have an AI note taker, you may have different points that come up live to help the salesperson navigate certain things. You may have like an AI sales engineer on the call as well. When do you think it will become more common where the actual salesperson is an AI?
11:59 Speaker 01 I think we're already seeing it. I've had lots of, not personal experiences, but I know lots of people who have that experience. I know so many other industries, home services, property management, whatever. People are already using voice AI, maybe not as a salesperson, but to solve other problems. That's happening right now. I just learned a fun story about Coinbase the other day, which is they have a voice AI avatar that screens all early career candidates that are coming through their top of funnel. And if you get past that, then you finally get to a human.
12:33 Auren Hoffman? And by the way, if I'm a candidate, I've got my own voice AI avatar that screens all the companies too.
12:39 Speaker 01 Yeah, yeah, yeah. So I wouldn't be surprised if that's already happening in sales in some way. I'm not surprised by the technology or when that's happening. I think I would just question whether that's right for the company or what trade offs they're going to face with that approach. Basically, I think the thing is when you're in voice AI for home services or plumbing or whatever, it's deterministic. There's a right answer and a wrong answer. Did you solve the problem or not?
13:01 Auren Hoffman? Often it's just scheduling too.
13:03 Speaker 01 Yeah, it's just like a deterministic path. That's not true in candidates and that's not true in sales. There's way more art in knowing the fit, both in candidates and sales. And I would question whether technology can do that now or in the near future.
13:16 Auren Hoffman? How is the SDR function changing? How have we seen it change today and how do you expect it to change over time?
13:23 Speaker 01 We're actually launching SDRs at Clay for the first time. We're creating our own function. We're seeing the best SDRs, all their manual work, all their manual research completely automated. So they're mostly spending their time on very high quality targeted individual outreach to like the biggest enterprises and multi-threading in very specialized way.
13:43 Auren Hoffman? Very custom.
13:45 Speaker 01 Very custom, very human touch or cold calling. Or they're using Clay and other tools to build their own systems to build pipeline. But they're not doing a lot of the normal SDR type work. A lot of that gets automated away and so you can be way leaner. At the moment, you still need a human touch. I've heard lots of blue chip unicorn companies that build internal systems or use Clay or whatever. And then when they added the SDR team, they saw dramatic improvements with the human element. But then they can actually take away a lot of that work and automate it. And then just keep the parts that are uniquely human that are uniquely creative.
14:18 Auren Hoffman? I'm super interested in like pricing and how that changes people's use of the product. And you guys opted to do some sort of usage based rather than a seat based pricing, which I generally like, I love air table, but because they have a seat based pricing and it's super expensive per seat. Almost no one at our company uses it. Basically, all of our stuff is still on Google Sheets. Whereas if they did a usage based pricing, we would use it for everyone and be using it all the time. We'd be paying them more.
14:49 Speaker 01 We started doing usage based pricing in like 2021 or early 2022. And it was very early in go to market. We may have been the first ones in go to market to do it because every time I met someone or a prospect, they were shocked that it was in seat based. And they almost didn't believe me. They thought they were getting the deal of the century. They were like, oh, everyone could just use this. For us, the insight was we don't actually want everyone to use it. Product is a product made for more technical audiences. By technically, I mean GTM operations or systems people or growth marketers that you're talking about. To build systems on behalf of a lot of people, we explicitly are not building for salespeople, actually. That's the intention. That wasn't a trade off in our mind. We actually want just a handful of people to build on behalf of other people.
15:31 Speaker 01 It was the value metric for us. At the beginning, we were more transacting data. So usage base was a better corollary for that than seat based, especially to cover our COGS. And obviously, AI has made that way more standard in the years since. It felt really natural and authentic to us, even though at the time it was unusual in our market.
15:51 Auren Hoffman? World of DAS isn't just a podcast. World of DAS membership is a private invite only community for founders and executives of important data and AI businesses. It's a place for off the record, high trust conversations focused on the biggest questions data leaders are facing. The community is about peer learning, curated intros and access to private events. Everyone in the group is building or leading something meaningful in data. If that sounds like you, apply to join at worldofdas.com. That's world of D A A S dot com.
16:32 Auren Hoffman? You've got a ton of different data providers. The most important thing about data is that the fact is true. And often with data, especially with marketing data, most of the facts in many cases are very high precision. Facts are not true or they're out of date or whatever. I'm sure with some data providers, you realize, oh, they're really good at this, or they're not good at this, or they're good in this geography, but not good in this geography. Or they're really good on cell phone numbers, but they're really bad on titles or how do you think about that?
17:00 Speaker 01 It's not clear cut and we spend a lot of time and energy on this. We try to publish these data tests that try to give more transparency to everyone else. But basically we do a lot of testing and there's match rates and fill rates, and then there's accuracy. We do our best and we run large statistically significant data tests on all of it.
17:17 Auren Hoffman? So you have some truth set somehow. You know that Varudh's cell phone numbers X or something.
17:24 Speaker 01 And there's a lot of sampling to really triple check the data. Then, by the way, we also have checks and balances. So we have actually a way in product for you to give feedback on the data to determine if it was right or wrong.
17:35 Auren Hoffman? Obviously, the email bounces, you'll know, but there might be other things within the product where people could just put in, "This is wrong," and that's like a feedback cycle.
17:42 Speaker 01 That's a feedback cycle to the provider, so there's always a continuous loop of improvement. And then, by the way, for like phone numbers and other contact providers, we have lots of verifications. So we have 10-plus contact verification sources that we partner with as well to track it and check that in real time.
17:58 Auren Hoffman? You're buying data from all these different providers, let's say about the contact information, about engineers or something. It could even be like 5% of the data is amazing, 50% is okay, and the other 45% is junk, and then you'll just use that 5% that's amazing?
18:13 Speaker 01 Yeah, yeah. For example, there are lots of providers that give you 100 data points, but maybe only two data points. Website traffic and revenue are really, really good. We'll just take the API calls for that and cut everything else.
18:26 Auren Hoffman? Or they're good for that, but just for France or whatever. They try to have it globally or whatever it might be. That makes a lot of sense. What we found for so many of these companies that sell data is it's very hard to get a premium on price when you're more accurate, because people, either they don't believe it, or they don't have a way of testing it or whatever. And so there's this price degradation. You can have more coverage, that's a way to get premium. It's very hard to actually prove that you're a better product. How would you advise data companies to help them prove they have a better product somehow?
19:04 Speaker 01 We'll just do a data test with you. If you want to prove that you have a better product, seriously, just tell us, because if you do, it's fully in our incentive to figure that out with you.
19:14 Auren Hoffman? And would you then like publish a blog on it?
19:16 Speaker 01 Yes, yes. And then it's in our incentive to tell the world that it's really good. So we're the distribution partner for all these data providers. And if you actually think your data is better, just tell us. Tell me. Email [email protected] . It's fully in my incentive to double-check that and make sure that's true, because if it is, you're going to make a lot of money and more people will come to Clay, so it's a win-win for everyone.
19:37 Auren Hoffman? One of the things with data is that it has a shelf life. If I have data about you today, you work for Clay, but back in time you didn't. Maybe five years from now, you might not, or something like that, or your title might change or whatever. It's changing quite a bit. It might be good today, but it might not be good in the future.
19:54 Speaker 01 You came to my website last night, but then that's useless to me if I wait two months to act on it.
19:58 Auren Hoffman? Correct. So how do you know from a data provider's perspective if they continue to be good? Are you retesting it in some ways? Are you re getting feedback?
20:07 Speaker 01 We do retest it. We do get lots of feedback from users. Generally speaking, data decays. I think it's like two to 4% a month. We're tracking that and monitoring it. But there are a bunch of data points that have longer shelf lives.
20:19 Auren Hoffman? The fact that your gender doesn't change as often or something.
20:22 Speaker 01 Mobile numbers aren't changing. Company addresses don't change that often. Putting rounds, unless you're a hot AI company, don't happen that often. So there are a lot of things that do endure. We obviously also push people to signals, things that are happening all the time and you're monitoring those. One of Canva's products is they help you keep your brand consistent across lots of different social channels. They use Clay to monitor LinkedIn and Twitter for posts from brands and then they use AI to analyze them and be like, hey, is this on brand or off brand? And if it's off brand, they like message the head of design and it's like, hey, this doesn't meet your brand guidelines. You should buy Canva. So there's lots of signals or like Vantus tracking companies that become SOC two compliant.
21:00 Auren Hoffman? So it's a go to market motion.
21:02 Speaker 01 Yeah, it's a go to market motion based on these signals.
21:05 Auren Hoffman? That's a very, very, very great email. If I got that even from Canva, I would be very excited if they told me something very specific about my company.
21:14 Speaker 01 That's what we try to focus on. You need to know when someone has a burning need for your product. What is the signal for your business? Is it they updated their support documentation? Is it that they're doing this thing that doesn't make sense online? Is it they just became SOC two compliant? Did they just get a new office and that's in the news? Let's find that signal because you know something about your customer that others don't know, and then you could build a system around it.
21:35 Auren Hoffman? More and more of our customers are doing research on LLMs, and we're seeing this move from SEO to GEO. How do you think that plays out?
21:44 Speaker 01 I don't feel like I have good insight on this question because right now feels very gamified.
21:48 Auren Hoffman? Just like SEO.
21:50 Speaker 01 I don't know about that. In a way, yes, but also in SEO, the cream of the crop typically tends to rise. And are there ways to gamify that? For sure. But typically, the higher quality gets to the top. With GEO, I don't know if that's clear yet. I also hear about crazy tactics that people are doing, and unclear if any of them work, actually. I've heard of a tactic where someone is just spinning up lots and lots of bots to ask the same question on these things.
22:13 Auren Hoffman? We've all heard those things for sure. We all have friends who've tried it, and I do think some of them have been working.
22:18 Speaker 01 Some of them work, some of them don't work.
22:20 Auren Hoffman? It's a cat and mouse game.
22:22 Speaker 01 It's unclear if any of this is making a difference, too.
22:25 Auren Hoffman? SEO is easier to measure. At least traditionally, when you and I have searched for best sushi place in San Francisco, we get the same results. Whereas when we're on ChatGPT and ask that question, we're often getting different results based on a whole bunch of different things that are happening.
22:41 Speaker 01 To be honest, my genuine reaction is I wouldn't change behavior too much based on this. I generally kind of think these moments happen in marketing all the time. There's always a new hot thing. There's always a new hot tactic. And this moment in time, maybe it's GEO. I don't think you're going to solve your problem with that by like getting the right tool for GEO or something. I think if you just make high-quality content that helps people, it'll work itself out. And make sure you distribute it in places that make sense, like don't put it in a black hole. I think if you just do that, I'm not saying do that and pray, but generally speaking, that's 98% of the work. And these optimizations on the side are necessary, but oftentimes people get really attracted by these fads, the hype, and then the FOMO.
23:26 Speaker 01 The FOMO is real because you're like, hey, someone's spinning up 6,000 bots to like do this. And you're like, fuck, why am I not doing that? But just do really good content that helps people, and it'll kind of work itself out.
23:36 Auren Hoffman? It's also interesting because these elements are changing so often. How they do things is changing. Seems like every few weeks they're changing.
23:46 Speaker 01 So how could you plan your business around that?
23:48 Auren Hoffman? I have a friend, he was getting a huge amount of traffic from chat GPT for his business. Who knows, they did some like random update and his traffic went down by forty percent. Literally one day had nothing changed, whereas Google, those updates happen way more slowly over time because it's just a more mature product.
24:06 Speaker 01 For sure.
24:07 Auren Hoffman? What else do you think is changing on B two B data? These intent signals are so important. I am not in the market for most B two B things very often. But when I am in the market or I'm getting close to the market, that cold email no longer is spam. That cold email is actually very valuable content that I want to see.
24:27 Speaker 01 Generally speaking, your ability to ask certain questions about your business and get answers to those questions is better than it's ever been before. I'll give you an example of this. We have a customer who's a logistics staffing company. They can only serve warehouses of a very specific size. But there's no good data on warehouses. So how do they do this? Well, they'd be screwed if this was three years ago. But right now, what they can do is they use clay to get Google Maps to show up all the warehouse locations. Then they use AI, then they pull the satellite images through Google Maps. Then they use AI to analyze the satellite images. With AI, they count the number of parking spots around the warehouse.
25:10 Speaker 01 And then they count the cars in the parking spots. And it turns out that that's the best predictor of warehouse size and employee count than anything else.
25:18 Auren Hoffman? By the way, I can send you later, but there are great companies that have square footage of warehouses.
25:24 Speaker 01 I'm just saying that for this company, they weren't able to find the data source that worked for them. And so they were able to use AI in this clever and creative way. I know companies that use AI to count chimneys.
25:35 Auren Hoffman? Solar panels.
25:36 Speaker 01 I know a company that does music licensing and does licensing for venues. But instead of having humans go and inspect all of them in person, they've AI analyzed the internet and reviews and Yelp to understand what genre of music they're playing, how loud the music gets, all these things to determine if they actually need the license or not in a way they could have never done before. The ability to ask crazy questions about you and your customers and whatever, and then answer them is better than ever before. That's really what's changing the market right now.
26:06 Auren Hoffman? You guys did a 10-year offer, $20 million 10-year offer, and you let employees sell some of their shares. Walk us through the logistics of how that worked and what you learned from it.
26:16 Speaker 01 This is the time we publicized it, but generally we try to offer employees a chance to sell whenever we do a fundraise, but the logistics are painful. We agreed to the terms with Sequoia, I think in January, maybe it was early February, and I don't think we were able to do it until like late April or May or something. It was an unbelievable amount of time with lawyers and stuff to set this all up. And by the way, it wasn't even $20 million. People didn't want to sell, so it was actually way less than that in terms of what actually got sold. My philosophy is give employees the ability to sell stock as often as you can, and there are legal constraints, whatever, but just as often as you can. But then also, the legal burden of a tender is quite high. And so I would try to just do one-offs with 10 employees at a time or something, and that's super fast.
26:59 Auren Hoffman? How does it work?
27:00 Speaker 01 I think if it's under 20 people or under 15 people, you don't need it to be a tender. You can just be like, hey, Orin wants to buy stock. Let's have the board approve it. Here are 15 people who want to sell stock. Go for it.
27:10 Auren Hoffman? And then you don't feel like as a company, you're like picking favorites as to who those people are?
27:16 Speaker 01 You'd have to logistically figure that out.
27:18 Auren Hoffman? Or do you just randomly pick them out of a hat?
27:21 Speaker 01 No, no, no, I'm just saying, I think you have to get interest in some way and do this in some fair way. I'm just saying the logistical burden of doing a tender, it took us three months and hundreds of thousands of dollars in legal bills and all these things. It's a lot.
27:33 Auren Hoffman? When you did the tender, you said you did it more than once. So when you did the second time, was it easier? Was there some sort of learning from the first time to make it easier?
27:41 Speaker 01 Well, I think it's more just when people want to sell and they tell us. We try to facilitate it if there's a way that we can do that. The formal tender, we will need it once. Generally speaking, it's like 10, 15, five people who are trying to sell or are interested. You can do that in a way that's not as logistically and legally burdensome.
27:58 Auren Hoffman? You could just say to the investor, hey, just go contact these eight people or something or whatever.
28:03 Speaker 01 I don't know all the details, but I'm pretty sure the company can coordinate it in a way that helps make it happen. And then you get the board approval and then it just kind of happens.
28:10 Auren Hoffman? Let's say you're doing a tender, you just raised money at a $3 billion valuation. You're doing a $50 million tender or something like that to your employees, but you don't want that to affect the 409A price of the stock. How does that work? Is it, if it's small enough, it's de minimis and it doesn't affect, or how does it work?
28:29 Speaker 01 That's generally the guidance we've got when we talk to our lawyers, that it shouldn't affect the 409A.
28:34 Auren Hoffman? So it has to be small enough to do it? Okay.
28:37 Speaker 01 Yeah, and generally you can make arguments with the 409A providers that feel compelling and reasonable, and they can work with you on that.
28:45 Auren Hoffman? If Sequoia is investing at a $3 billion valuation, but they're getting this preferred stock that does have a lot more value than common stock, are you doing the tender at like some sort of discount to that? Or do you convert it into common when Sequoia buys it?
29:01 Speaker 01 We don't convert, it's just all common if it's a tender. But basically it just depends on the market demand for your stock. If you are a really good company that's doing really well, investors will pay the same as they would for preferred. If you are not a good company, they're going to ask for a discount. I think that's just the market realities.
29:18 Auren Hoffman? Preferred does have preferred rights. It should have some primae.
29:22 Speaker 01 That's the thing. Play, for example, we're really efficient as a business. We don't earn that much money at all. We don't need to raise that much money. So, hey, let's just say we raised $100 million, but there's a lot more demand. People want to invest more in the company. We just don't need more primary. In fact, we didn't even need the 100 million for the record. But they're like, hey, you can buy secondary if you want. And the market demand is so high that they're happy to pay the same price.
29:45 Auren Hoffman? It either means that the preferred is too cheap, or it means the common is too expensive. One of those two is because the preferred is better. So it does mean if you're getting them at the exact same price at roughly the same time, the preferred is potentially too cheap then.
30:00 Speaker 01 I don't think the preferred has to be too cheap, but I also think that the rights of preferred shareholders gets less valuable as the company gets more de-risked.
30:07 Auren Hoffman? But it still has some value.
30:09 Speaker 01 It has some value, but maybe you say, hey, I can't get access to this company that I'm trying to invest in, and so I don't care about that access. I don't care about those rights.
30:19 Auren Hoffman? Yeah, that makes sense. Obviously, like you see rounds done at $3 billion and then all of a sudden someone can't get in. And they're like, well, I'll give you $4 billion to get in. So the same thing happens all the time. So this is just kind of a version of that. You guys also did like a community round. What is that?
30:34 Speaker 01 A community round is basically a fundraising round, but for your community, our customers, and we have a big community. So we have 70 clay clubs all around the world. We have agency partners that are making millions of dollars a year built on clay. We have a lot of clay supporters and honestly, we are a community driven company. We are building for the community and they are why we've been successful. We wanted to find a way to have them be a part of our story and feel like financially and economically bought into the company. Because they're the ones helping build it. They're the ones getting economic benefit too from it. This was an opportunity for us to do that.
31:07 Auren Hoffman? How does it work?
31:08 Speaker 01 Another logistical mess, I would say.
31:10 Auren Hoffman? Some of these might not be qualified investors and...
31:13 Speaker 01 The thing with this is you have a few options. There are a few providers that do this. We used a company called Republic. You could maybe consider Angel List, theirs, WeFunder as an option.
31:24 Auren Hoffman? And you're creating essentially like a R U V type of instrument?
31:29 Speaker 01 There are trade offs with all of these things. With Angel List, everyone has to be an accredited investor, which is like a high bar for people in our communities. They're not really set up for this. They're not really set up for like hundreds, if not thousands of people to participate in this. WeFunder, the details were a little hazy because this was like a year ago, but WeFunder had this flow that didn't make sense. Kind of works like a GoFundMe where they're like, hey, put up this page, first come, first served. And I was like, no, no, no, no, no. What I want is a setup where you say, hey, we're doing this. You should apply and then we select the people.
31:59 Auren Hoffman? Or you might want to change their allocation. They want 30 grand of allocation. You're like, sorry, you can only get 15 in the end or something.
32:05 Speaker 01 Exactly. So we want all that flexibility. And then the other thing that WeFunder asked is they were like, hey, you should need to reserve some of your allocation for WeFunder VIPs, random people who we don't know. And I'm like, I don't care about them. I want only a clay community. So that's why we didn't do WeFunder. Third option was Republic was mostly okay. And maybe they've improved this since then, but the logistics of actually doing the funding, we had a big drop off of people who got accepted and then actually invested. Because the logistics of doing it were very cumbersome. It was like a very cumbersome, confusing process on the Republic site.
32:40 Auren Hoffman? One of the reasons I love Angel List is if you're on it, you literally press a button and the money just goes.
32:46 Speaker 01 No, no, no. It was way more complex than that. And so we had a real drop off, which was unexpected because we didn't diligence that part of the process that well.
32:53 Auren Hoffman? I always wondered like why more people don't do it. And it sounds like more people don't do it. It's just freaking annoying.
32:57 Speaker 01 It's really annoying.
32:59 Auren Hoffman? A few personal questions. You cold emailed your way to a job with Hillary Clinton. First of all, like walk us through that story and then I have a couple of questions about cold emails.
33:09 Speaker 01 I also cold emailed my way into my job at Google after that. And I cold emailed Kareem, my co-founder at Clay, which is how we met.
33:17 Auren Hoffman? I love this. By the way, I've met many friends through cold emails.
33:21 Speaker 01 And by the way, you cold emailed me.
33:22 Auren Hoffman? Exactly.
33:23 Speaker 01 Was that a real email or was that automated email?
33:25 Auren Hoffman? That was a real email.
33:27 Speaker 01 It was hard to tell.
33:28 Auren Hoffman? A real email that was templatized.
33:30 Speaker 01 Okay, it was templatized. Yeah. So Hillary, it's a good story and I'll tell the story with Kareem too, because that's also really good. I was 18 years old. I was obsessed with politics. I thought I would do that for the rest of my life. I was a freshman at Penn and I really wanted to work for Hillary Clinton in particular, early 2013 or something. She had just left the State Department. I was trying to find a way to get in touch with her and work with her in some way. I was actually just really passionate about her. I'd admired her for a long time. I just wanted to be a part of working with her in some way, shape or form. At the time, she was just a private citizen, actually, but I was just like, I want to be involved. And it turns out that she had a small personal office of like six people. I hunted down the senior advisor to her, his personal email. When you're a student, you have access to alumni databases.
34:14 Speaker 01 The email tools back then weren't as good as they are now. So I was at Penn, but this guy went to Columbia. So I had friends at Columbia and they went through their alumni database and got me his email. So I got his email that way and then I cold emailed him 10 times. And he didn't respond. And then the 10th time, he's like, okay, fine. I'll meet you. But just let me know the next time you're in DC. So obviously the next week I invented an excuse to come to DC. I'm in Philadelphia, very easy and Bolt bus for the win. He was like, okay, fine. Meet me at the Starbucks below my office. So I'm like, okay, I come to the Starbucks. What I didn't know at the time is he tells his assistant, who I'm now friends with. Hey, this crazy person has emailed me 10 times and has asked for a meeting. Just interrupt the meeting in five minutes and get me out of this.
34:57 Speaker 01 So I meet with him for five minutes. The woman comes down, interrupts the meeting, says, Philippe, you have to go. But the first five minutes went so well that she ends up just staying for the next hour. And then he more or less offers me a job at the end to be like the seventh person, the intern, the junior person of this personal office.
35:13 Auren Hoffman? That's awesome. I love that.
35:14 Speaker 01 And then I worked for her throughout college. I would basically commute from Philadelphia to New York and DC. Very convenient again during the year, then spent all my summers working with her. And then graduated a year early in 2015 to kind of work for a campaign full time and did a variety of roles with her there. And then with Kareem, Kareem originally started Clay with his best friend from college, Nikolai, 2017. They didn't find a product they were excited about, so there was a lot of rotating and pivoting for like three or four years. And they didn't really commit to a particular vision. I was starting another company at the time and was in the no-code space. I was in a Slack group where someone mentioned a webinar about Clay. But at the time, Clay didn't have any revenue, no customers, no users or anything, just like a very early product. They were trying to do a webinar.
35:58 Speaker 01 So I sign up for this webinar, but no one else shows up. I'm the only one who shows up. And it's not even Kareem, it's someone else. I started using the product, playing with it, and I'm like, whoa, this is really crazy. I'm really into this. I couldn't stop thinking about it. So I cold emailed Kareem, but he doesn't respond to me because he's on a silent meditation retreat. So I cold emailed Nikolai and he does respond to me and we hit it off. I'd like to think, actually, I told Kareem since this, I think if Kareem responded, none of this would have happened. Because Nikolai and I have very similar personalities, but Kareem and I don't. So I don't think we would have hit it off in the same way. And Kareem is like inherently suspicious of things like this, whereas Nikolai embraces it. So anyways, I hit it off with Nikolai, we spend more time together, we start working together. But then Nikolai decides to leave the company to pursue his own things. And a few other people in the company leave, which is very bizarre. So the company resets in a lot of ways.
36:43 Speaker 01 It was just Kareem, one or two engineers. And then I joined as a late co-founder for whatever reason. And we're like, let's just commit to this new vision, let's reset the business, let's launch the product in this new way and focus on go-to-market and went from there.
36:56 Auren Hoffman? That's awesome. I love that. If you're advising people to write a good cold email, how would you advise them?
37:02 Speaker 01 I get a lot of cold emails. Most are bad. There's probably a couple of things you're trying to do. There's a lot of advice where people are like, be as brief as possible. But I've actually gotten cold emails that are unbelievably long, in fact, crazy long, but also worked. So I don't think that advice on just write super short emails that are to the point. People always mention that Snapchat email, you know, to Evan Spiegel. That was like four sentences and to the point. I've actually had emails come to me that are paragraphs upon paragraphs, essays. But are so good and compelling that I'm like, okay, someone took all this effort to write this. Let's talk. Basically, you're trying to convince someone that you can do the job before you've done the job. You're trying to solve problems they have. You're trying to communicate something that they should give you a shock.
37:47 Speaker 01 And so what could that be? There are many options. There's not a script or a template for this. There's a lot of options. You could have some insight about their product, about their company. You could do a lot of research and be like, hey, here are five problems that I found and here are fixes for all of them. That's me doing my job before I get there, before you've even met me. That's cool. I would meet with someone who did that.
38:08 Auren Hoffman? I've used your product and here's how the product can be better. Here's what I like.
38:12 Speaker 01 Yeah, or just here's on your website, here are five issues. I just got a cold email from a team in India that makes clay nation videos. And they made a custom clay nation video for clay that was handmade. They made all the clay themselves and they showed a lot of effort in doing it. So I watched the video, I'm like, wow, this is a lot of effort that they've gone to. Now I'm engaging. Now I'm introducing them to our brand. And they went to a lot of effort just for a cold email that they have no idea what's going to happen.
38:39 Auren Hoffman? Those are the ones that are going to convert well. If you're passionate about it, it's all these other things that go through it.
38:44 Speaker 01 So I think you're just trying to communicate something like that. And that's what's worked when people cold email me.
38:49 Auren Hoffman? How are you personally using AI tools outside of clay?
38:55 Speaker 01 There's a lot of ways. One is Whisperflow. I use Whisperflow all the time. This is like a voice to text product all the time. I'm using Whisperflow.
39:03 Auren Hoffman? And I use Whisperflow on my desktop, but is there a mobile version?
39:07 Speaker 01 No, there's mobile too. Yeah, yeah, yeah. I use it on mobile.
39:09 Auren Hoffman? I should start using it.
39:11 Speaker 01 It just works and I do a lot of text through that. I'm a big audio note guy. It just works. The other thing is a friend of mine built a product, which he then pivoted to something else in property management AI. But he built a product for me, basically just me, because I think I'm the only customer that auto sorts my inbox into different categories using AI. And it's amazing. So there's just three categories now and it's just action, just read or archive this. And it saves me unbelievable amounts of time because I get hundreds of emails a day.
39:40 Auren Hoffman? I have like 400 filters that put them into the read folder. This is like automatically does it. So you don't have to do all those filters and stuff.
39:48 Speaker 01 Yeah, just three things and it's almost always right. It saves me an unbelievable amount of time.
39:53 Auren Hoffman? That's so cool. All right. Two questions we ask all of our guests. What conspiracy theory do you believe?
39:58 Speaker 01 One is the Denver airport, which is bigger than the city of San Francisco and was dramatically over budget and dramatically over timeline. And there's a lot of conspiracy theories about what happened there. Is it a secret bunker? Is the government using it? Is there this underground town happening there?
40:12 Auren Hoffman? Why would there not be like a huge underground system if it's that big? You could fit a lot of stuff under it. It makes sense to go do something there.
40:20 Speaker 01 So I think I believe that. And then another one that I asked a friend about is shark bites. He was like, there's only 70 to 80 reported shark bites a year, but lots of people know individuals who've gotten bitten by sharks and the numbers don't add up. And so the conspiracy is that towns and papers don't report the shark bites because they don't want it to be disincentive for tourism.
40:43 Auren Hoffman? God, so people won't go to like Miami if they think all these people are getting bit by sharks. But if it's, oh, you have less likelihood of getting bit by a shark than you do by hitting by lightning or something, then okay, great, then it's okay.
40:57 Speaker 01 But sharks are out there and they're dangerous.
40:58 Auren Hoffman? That makes sense. There's no incentive to advertise. Maybe if I'm a competing area or something, I could advertise that Miami gets all these shark things because no one in my area gets shark stuff or something.
41:10 Speaker 01 Exactly.
41:11 Auren Hoffman? That's super interesting. Our last question we ask all of our guests, what conventional wisdom or advice do you think is generally bad advice?
41:19 Speaker 01 Honestly, I have like a hundred of these.
41:20 Auren Hoffman? Awesome. Let's dive in.
41:22 Speaker 01 We could be talking for hours about this because I feel like at Clay, we just do a lot of things somewhat uniquely or authentically to us that always buck conventional wisdom in some way, shape or form. I'll try to just do a couple so we can stay focused here. One is, and this is kind of a little bit of a comment about 996 culture, but there's conventional wisdom, especially in early stage companies, but just generally the people should just always be working crazy hard and working 996, working really hard all the time and grinding. And then when they're like, oh, you don't need to do that in a mid-stage company because you have lots of people, you have product market fit and you need to optimize burnout. I'm actually thinking of it differently. And they're like, in an early stage company, you need it the most because you need to stay ahead.
42:04 Speaker 01 But what I find, actually, this insight was inspired by Kareem. What we found is that at early-stage companies, you're oftentimes constrained not by time, but you're constrained by insight. Because you need an insight to truly find that product-market fit. You're not usually just in this random rat race.
42:22 Auren Hoffman? That's why he was at that silent meditation retreat when you emailed him.
42:25 Speaker 01 That's right. You're really constrained not by time and your ability to just do a million things. You're not building like a packaging business or something usually. You're really constrained by an insight that will help you truly differentiate and find unique product-market fit that could help you build this extraordinary company. Because that's how the best companies are usually made. It's through a different insight, not just doing the same thing everyone else is doing, but a little bit better. I think that's true and something we share often. Another thing is around interviewing. I think there's a lot of pressure and wisdom around you should interview lots of people, basically, and you should just meet lots of people. I feel this FOMO all the time when I meet other founders or executives and they're like, oh, I interviewed 100 candidates for this role. A lot of the times, I hire the first person I meet or the first person I liked. And I feel pressure about that.
43:10 Speaker 01 I feel a lot of insecurity about that.
43:12 Auren Hoffman? If you thought someone was good for the job, why would you not hire them?
43:16 Speaker 01 Well, you feel insecurity because you're like, am I doing the best job? Did I do a good job? Am I doing my fiduciary responsibility to find the best fit for the company? I think the truth is you don't actually have to be exhaustive. You just have to be right. It just has to feel right. And a lot of times when you're at companies like Clay, good people come around. And by the way, this is a pivot to another piece of conventional advice. When you're at companies like Clay, VCs, lots of people inundate you with experienced hires. I think if you look at our org and you just remove the EPD side of the company for a moment and you look at the business hires, half of those people at Clay have never done their job before.
43:50 Auren Hoffman? I'm surprised it's not like 90% or something.
43:53 Speaker 01 There are some people who know what they're doing, but they're coming from completely different backgrounds. It's archaeologists being customer success people. It's physicists being events people. And that's what creates the special sauce, the creativity. Politics, people running ops. I'm not saying experience doesn't have a place. Certainly, if you're hiring a lawyer or an accountant or a data scientist or an engineer, you probably should hire someone who knows what But I think it just comes at the right fit and the stage. Our salespeople aren't salespeople. We don't hire salespeople to do sales. We don't pay commission. And only now are we hiring more seasoned salespeople because that's a better fit for us now. So it's just thinking about the fit. And a lot of times people feel pressured to just be like, let's hire the person who's done it before. Whereas I know there's always advice, don't hire someone at Microsoft to do something at Figma or whatever.
44:38 Speaker 01 But people won't make that obvious mistake. But now people are still making the mistake of hiring someone who's done it at a startup to expect them to do it again at another startup. And even that, I would say is flawed in many ways. And truly is that person really the right fit for your stage and company, what you need?
44:55 Auren Hoffman? Now, you mentioned your own phone. Do you think this person's good? You don't have to interview a hundred people. You could hire them. You're also married. How did you do that with your spouse? Did you have to go through 18 people first? How'd you make the decision to get married there?
45:08 Speaker 01 No, I think the same thing happened with my wife. I don't want to say she's the first relationship I was in, but like the second or the third. I met my wife when I was 23. She was actually six years older than me. She was 29 at the time. But I was very young when I met my wife and just committed. It felt good. I was happy. I am happy. It's working. I'm not trying to overthink my happiness. It's working. I'm happy. Everything's good. It's again, easy to feel FOMO or pressure of is the grass greener? Is this happening? I'm happy. Things feel great. We're very aligned. We're in love. Don't overthink that because most people aren't even in that state.
45:40 Auren Hoffman? If you're happy and love, they like you. That's about all you could ask for. That's pretty amazing. This has been amazing. Thank you, Ruud, for joining us on World of DAS. I follow you at VXAnand on X. I definitely encourage our listeners to engage you there. This has been a ton of fun and super interesting.
45:57 Speaker 01 And LinkedIn, that's where we're pushing the content, LinkedIn.
46:00 Auren Hoffman? I also follow you on LinkedIn as well. So yeah, absolutely. This has been amazing. Thank you again.
46:05 Speaker 01 Awesome. Thanks, Auren.
46:09 Auren Hoffman? Thanks for listening. If you haven't already done so, please subscribe to World of DAS wherever you consume your podcasts. YouTube, Apple, Spotify and more. And please help us get discovered by leaving a review and check out worldofdas.com. That's worldofdas, D-A-A-S, dot com. And of course, connect with me on Twitter, at Auren. That's A-U-R-E-N. Would absolutely love to hear from you.
Previous The LM Brief: How Leaders Can Close the AI Intelligence Gap Next The LM Brief: The Energy Bottleneck Behind AI’s Growth Page updated November 1, 2026 · transcript and summary are machine-generated