Varun Anand

The through-line

Varun Anand keeps returning to one idea: the binding constraint in building a company is creativity and insight, not effort. Asked what founders should optimise for when they lack product-market fit, his answer is that "what you are constrained by is uh your insight" and "what you are not constrained by is time," which makes the online culture of grinding "counterproductive and a little deceiving" [2]. Hard work buys more cycles, but it does not manufacture the way of thinking that separates you from a competitor. The same conviction runs through everything else he says. It is why the number one thing he screens for in candidates is creativity [1], why he thinks B2B marketing fails, why he believes the traditional sales org chart is broken, and why he describes the ceiling on his own product in terms of the user rather than the software: "the limit to clay as is to notion is is your creativity" [2].

The second thread is that structure should follow that creativity rather than constrain it. He cites Ken Stanley's *Greatness Cannot Be Planned*, whose argument he summarises as following interestingness and then having "the the skill to acknowledge it for what it is and then double down and go for it and hustle and execute," and notes that goal setting is "famous for stifling our creativity" [2]. Clay's own history is that story: years of exploration and pivots before a commitment to go-to-market, and a product whose whole pitch is giving teams somewhere to try new ideas fast [2][3][4].

On what actually grows a business

He reduces growth to a small set of moves: everyone has a TAM and is "just trying to find the delta," and the only routes are converting inbound, going out and getting outbound, or expanding existing customers [1]. Inbound and outbound he treats as the same mechanic differing only in "where's your lead originating from" [1]. Clay's role is to supply the raw material for those moves: every data provider in the world, LLMs to improve that data, and data points nobody would otherwise hold, such as whether a company is SOC 2 compliant or whether it sells B2B, used as "primitives to run growth campaigns" [1]. From there the vehicle can be personalised ads, outbound, or direct mail [1].

The mechanics he demonstrates are a single workflow: identify who signed up, enrich them, qualify them, score them, then generate the outreach language and move quickly, the same shape of workflow he says Anthropic, Figma and Cursor run [1]. AI columns classify job titles into CXO, VP, director or head; company size and sales headcount qualify the account; website scrapes surface fit signals, such as a benefits company checking what benefits a prospect already offers [1]. Email coverage comes from waterfalling across roughly 150 data vendors, trying the next one when the first returns nothing and validating the result, which is what produces the highest coverage [1]. Beyond the core use case of finding customers, he has seen people build websites programmatically at scale, run entire companies through Clay, generate tarot card readings, and find OBGYNs licensed in particular ways [2].

On why B2B marketing is boring and brand is not a luxury

His blunt verdict is that "B2B marketing is very boring extremely boring" [1], largely because marketers imagine buyers as Gartner and Forrester readers when a CRO "scrolls LinkedIn just like the rest of us" [1]. Brand at Clay predates revenue: they bought clay.com before any revenue, built a fifteen-person brand team, and contracted a clay artist whose blobs came with their own subscription model, roughly $3,000 a year with an annual renewal, which he happily paid for [1]. His justification is authenticity rather than tactics: "You wouldn't hire the head of brand or have the paymation artist or pay this much money from blobs every year" if it were not core to you, and over time it turned into a real differentiator [1].

He is candid that creative work has a hit rate. "Some are hits and some are flops. But when you're being this creative that that's going to happen" [1]. What he claims unequivocally is mind share, "the hardest thing to do in B2B marketing," in a world where attention is contested by every company and by TikTok and the news [1]. The evidence he offers is anecdotal but consistent: a month in Europe where nearly everyone he met had seen the videos he and his CFO make analysing marketing ROI, including one about wanting to take five prospects to the Super Bowl for $60,000 and being told no [1]. Mind share then shows up downstream as accelerated deals, new inbound, and being top of mind at the moment a buyer is ready [1]. He attributes the ability to do this to hiring "very non-traditional people to do marketing" [1].

On the GTM engineer and the end of the sales assembly line

The role came out of a practical problem. Three years ago Clay was hard to understand and hard to use, so instead of traditional sellers he looked for "people like me who are like growth and ops people, people who are like our customers," and the profile worked; the sales team is now around 35 people, still hiring that mould alongside some traditional AEs [1][2]. He and Kareem settled on the name GTM engineer, distinguishing it from a forward deployed engineer, which he treats as closer to professional services delivered on site [1]. He is careful about generalising: "no other company can do this actually because it only makes sense for us because our product is clay," and Figma would not have designers sell Figma [1][2].

The role outgrew Clay's own hiring because customers began hiring the same profile to run Clay internally [1]. His definition is people building revenue systems with automation and AI, running growth experiments, sometimes writing their own software, all "operating with a product and growth mindset" and iterating fast [1]. The structural argument is that the classic chart, SDRs prospecting, AEs closing, RevOps maintaining systems, marketing creating pipeline, "operates a lot like an assembly" line, and "assembly lines aren't really good for sharing learnings and and scaling ideas" [1][4]. The replacement is a small centre of GTM engineers who absorb the manual research and manual work on behalf of every rep and build systems that scale the best ideas the sales team surfaces [1][4]. Clay has since launched a go-to-market engineering marketplace matching its community to companies hiring for the role, naming cursor, framework, web flow and openai among them [1]. He also sees marketers becoming far more technical, and pushes back on framing that as science: what he observes is "the engineering discipline being brought to marketing," a systems mindset that builds repeatable pipeline machinery, where five years ago good marketers were Zapier masters and today reach for tools like pipedream [4].

On hiring, take-homes and work trials

Creativity is the top thing he evaluates in GTM engineer candidates, and it gets demonstrated in the take-home rather than asserted in an interview: a typical brief is three companies and three creative outbound campaigns built out for them [1]. Clay runs take-homes for every role, generally uncompensated as part of the process, with compensation reserved for unusually long ones [1]. Early on they ran work trials for almost every role; now they reserve them for roles that are hard to evaluate through a traditional process, and he gives the example of a colleague who spent two or three weeks on a work trial before her actual first day [1]. Underneath the mechanics, his stated hiring philosophy is about understanding human attributes and values such as low ego rather than checking qualification boxes [1].

On finding product-market fit and the reverse demo

Clay's origin was not a go-to-market idea. The question was "how can we um bring the power of programming to an order of magnitude more people," which led to the spreadsheet as the medium, and then to the observation that go-to-market teams already live in spreadsheets [2]. That eventually became a mission he now states as helping grow businesses and bringing the power of creativity to more teams [2]. Getting there took years of exploration by Kareem and Nikolai after 2017, during which they built things resembling Retool and warp.dev before those companies existed, without committing to a vision [2][3]. Anand arrived through a webinar he was the only attendee of, insisted it go ahead rather than be cancelled, and afterwards could not stop thinking about the product [2][3]. The company reset after Nikolai and others left, and he joined as a late co-founder to commit to the go-to-market vision and relaunch [3].

The method for reaching product-market fit is one he thinks others can copy. Running a sales-led motion for a $200 a month product is, as he acknowledges everyone told him, unprofitable and unscalable, but he and Kareem "almost irrationally believed" they could convert it to self-serve [2]. The tool was the reverse demo: ask the prospect to arrive with a real use case and some data they want enriched, share their screen, and aim to "blow" their mind by solving the problem in the first twenty to twenty-five minutes [2]. Every small friction got logged and fed to the team. He ran this eight times a day, five days a week, and over nine to twelve months the requirement went from seven demos to close a $200 a month deal to no demos at all [2]. His summary of why it worked is compounding: "it was just this constant iteration of taking the product feedback eight times a day, five times a week... and it compounds over time" [2].

On layering enterprise sales on top of self-serve

He would not touch enterprise until self-serve felt solid, which he defines concretely: creating demand, converting it on the website, activating users in the product, monetising effectively, and expanding through a value metric that makes sense in pricing [2]. The second condition was a person he trusted to scale it, in this case Bruno, hired as head of growth marketing and now head of marketing [2]. The reason to add sales was simple ambition: "if we want to be a generational company, we need other generational companies to use our product," and those companies would not adopt without help [2].

The path there involved two documented mistakes. Doing the work for customers as professional services put Clay in competition with its own agency network, hundreds of agencies, many earning more than a million dollars a year, which he credits as a contributing factor to Clay's product-market fit, and it was an unscalable amount of work at that price point [2]. Then a platform fee with usage credits on top proved brittle, because procurement negotiations tear a platform fee apart and expansion was hard [2]. They consolidated into single credit-based pricing modelled on Snowflake's, which aligned them with customers and scaled better [2]. The wedge is deliberate: focus on data enrichment, where "We are best in the world at that truly," because everyone has the pain of too many vendors and poor quality, it is quantifiably provable through a data test, there is an existing budget line item to compete against, and you can start without ripping anything out and consolidate over time [2]. Once trust and a contract exist, hackathons take customers to the more creative workflows that are the actual dream [2]. He also reframes selling in a way he can live with emotionally: less selling, more "let me help you find your fit," whether that fit is the self-serve plan or something priced to their needs [2].

On why a consultative, non-commissioned seller wins differently

Because Clay's sellers are GTM engineers rather than salespeople, the sales dynamic changes: "it's just that if you're being sold to by a seller who's incentivized in certain ways, it's just a different dynamic than if you are sitting across the table from someone who is not financially incentivized to close you, who is actually your who has done your role before" [2]. He is explicit about the trade-off using a framework from Dennis Leandreas, long-time CRO of Procore: a sale needs four wins, political, technical, legal and commercial [2]. Clay's team is naturally strong on the technical, legal and pricing wins, and weaker on the political win, which maps to traditional skills such as champion-based selling, MEDDIC-style frameworks and executive discovery calls [2]. His judgement is that the trade is worth it, because the builder relationship creates long-term trust with a good foundation [2].

On the CRM losing its centre of gravity

He reads the go-to-market market as unusually lopsided: Salesforce and HubSpot capture enormous value as systems of record, and beneath them sit point solutions rather than other major businesses [4]. His explanation is that many of those companies were founded by salespeople who productised something that worked for them, which makes sense but ages badly, because in go-to-market "you kind of need to keep changing," and once a tactic is copied you must move on and the product no longer fits [4]. What the market needs instead is what he calls an IDE, "where you can uh come up with new ideas and iterate really quickly, experiment really quickly, and execute them really quickly," and he thinks Clay is that place [4]. He is direct about the consequence: point solutions that simply flag a job change or a website visit, naming Champify and UserGems as examples, "are not going to endure in this transformation" and will become features of a larger platform [4]. Individual data providers, by contrast, will endure, because niche data points will always require someone to go and get them [4].

Alongside that, he sees the CRM itself demoted. The most cutting-edge teams "actually not rely on CRM," using Snowflake or Databricks as a flexible store of data, orchestrating with something like Clay, so "the data just finds the rep at the moment that they need it" [4]. That might be a text before a meeting, a Notion doc briefing before meeting a prospect, or a Slack ping, and it points to a world where reps use UIs less [4]. Clay runs this internally: an automated post-meeting follow-up he can edit or approve, and a morning briefing of overnight signals with suggested actions, all delivered in Slack and email rather than in a new platform [4]. He does not exempt the incumbents from criticism, granting that HubSpot is far easier to use than Salesforce while noting that both hold entrenched ecosystems and moats that are starting to shift [4].

On the data that does not exist yet

Asked what data is missing, he points first to businesses that are still effectively offline. Coverage for anything online is good, but "offline local businesses, small businesses outside of the US uh still doesn't have like remotely as the kind of data" that people take for granted with tech companies [4]. He extends this to the US as well, arguing the evidence is the wave of search funds rolling up car washes and mechanics, which he doubts would be happening if the data were up to par, and that consumer internet penetration numbers mask how far behind businesses are [4]. The second gap is closed ecosystems such as Meta, where it remains "remarkably hard" to build a custom audience and advertise against a small set of, say, 5,000 people, and making that easy is something Clay is actively working on, whether the endpoint is advertising or personalised cookie jars [4].

LinkedIn is the recurring illustration. He credits it with an extraordinary trove of data it would probably admit it has under-exploited from a product perspective, and notes that hundreds of thousands of companies sprang up in that absence, met with litigation over terms of service [4]. His reading of the legal position is that federal rulings allow scraping of public data, however LinkedIn feels about it [4]. On agents acting on a user's behalf he is more circumspect, calling the area confusing and messy and saying Clay is watching it closely, while acknowledging platforms are likely tracking mass activity [4].

On where AI does and does not replace the human

He does not think an AI closing deals is a technology question. Voice AI is already routine in home services and property management, and he cites Coinbase using a voice AI avatar to screen early career candidates at the top of funnel before any human involvement [4]. His scepticism is about fit rather than feasibility: those working cases are deterministic, with a right and a wrong answer and often just scheduling, whereas "There's way more art in knowing the fit both in candidates and sales," and he questions whether the technology can do that now or soon [4]. What he does see is the SDR role being hollowed out on the manual side. Clay is launching its own SDR function for the first time, and the pattern with the best SDRs is that manual research and manual work are completely automated so their time goes to high-quality, targeted individual outreach [4]. This is consistent with the broader shift he observes, where headcount patterns in marketing change shape as work is centralised into systems rather than distributed across reps [4]. He has also discussed publicly why go-to-market AI is a different problem from AI elsewhere, and why the assumptions that held in 2022 no longer apply [6].

On cold email as the demonstration of effort

Three of the pivotal moves in his career came from cold emails: into Hillary Clinton's small personal office, into Google, and into Clay [3]. The Clinton story is his own template. At 18 he tracked down the personal email of a senior adviser by getting a friend to search a Columbia alumni database, emailed ten times without a reply, got a grudging offer to meet next time he was in DC, invented a reason to be there the following week, and turned a five-minute Starbucks meeting, which the adviser had pre-arranged to have interrupted, into an hour and a job offer [3]. With Clay, the cold email to Kareem went unanswered because he was on a silent meditation retreat, so he emailed Nikolai instead, and he now believes the accident was fortunate, since he and Nikolai had similar personalities while Kareem "is inherently suspicious of things like this whereas Nikolai embraces it" [3].

His advice rejects the usual formula. He receives many cold emails and "Most are bad," and while everyone cites the four-sentence email to Evan Spiegel, he has been won over by messages running paragraphs upon paragraphs "so good and compelling" that he took the meeting purely because of the effort [3]. The principle underneath is that "you're trying to convince someone that you can do the job before you've done the job" [3]. There is no script; there are options, such as an insight about the product or company, or a list of five problems you found on their site with fixes attached, which he characterises as doing the job before you arrive [3]. His favourite recent example is a team in India that handmade a claymation video for Clay, making all the clay themselves, for a cold email with no guaranteed outcome, which earned them a viewing and introductions [3].

Takeaways

• The scarce resource in pre-product-market-fit companies is insight, not hours: "what you are constrained by is uh your insight... what you are not constrained by is time," and the grind narrative is "counterproductive and a little deceiving" [2].
• Convert a sales-led motion to self-serve by running reverse demos at volume, solving a real use case live in 20 to 25 minutes and feeding every friction point back to the team; eight demos a day, five days a week, took Clay from seven demos per $200/month sale to zero [2].
• Do not add enterprise sales until self-serve reliably creates demand, converts, activates, monetises and expands, and until someone trusted is in place to scale it [2].
• Professional services backfired by competing with Clay's own agency network, and platform fees got shredded by procurement; a single credit-based model inspired by Snowflake aligned incentives and scaled [2].
• Sell through a narrow, provable wedge, data enrichment against an existing budget line item, then expand into creative workflows once trust and a contract exist [2].
• The sales assembly line of SDR, AE and RevOps blocks the sharing of learnings; replace it with a small core of GTM engineers who automate manual research for everyone and scale the best ideas [1][4].
• Screen for creativity with a real take-home, such as three creative outbound campaigns built for three named companies, and use work trials for roles that traditional interviews cannot assess [1].
• Brand spending before revenue is defensible only when it is authentic: Clay bought clay.com and paid for subscription clay sculptures pre-revenue because it was "core to us," and the payoff is mind share, "the hardest thing to do in B2B marketing" [1].
• Tactical point solutions that flag job changes or site visits will become features of larger platforms; individual niche data providers will endure [4].
• The cutting-edge stack replaces CRM dependence with Snowflake or Databricks plus orchestration, so data reaches the rep in Slack, email or a briefing doc rather than a UI they must visit [4].
• A cold email's job is to prove you can do the job before you have it; length matters far less than demonstrated effort, research or fixes to problems the recipient actually has [3].