Clay's users make most of the content. Its executives post to a plan. A growth team runs a livestream series that has to pay for itself, hands sellers a list every Monday, and spends half its time selling the content inside the company. Clay says more than half of its pipeline now traces back to marketing, with no cold outbound. Here is how the machine fits together, in Clay's own numbers.
The Content Machine is a new column, and a different one from Series Teardown. That column takes one unusual content series and works out why it lands. This one is about the whole operation. Each instalment takes a startup or scaleup that runs content as a system and goes through what it makes, how often, how it reuses it, where it goes, who makes it, how it is measured and what it does for the business. The rule for getting in is that the company has said in public, with numbers, that the content moves revenue. Reach on its own does not qualify. Clay is first because it has published the clearest account I have read of how a content machine fits together, and because the numbers in that account are the company's own and I can label them as such.
The premise
Clay sells a data and automation tool for sales and marketing teams. You load a list of companies or people, it enriches the list from many data sources, writes outreach with AI and runs the plays. In September 2026 the company announced a US$115 million Series D at a US$7.1 billion valuation and said more than 17,000 teams build on it.
The machine, in one paragraph. Marketing builds programmes that get users to make content about Clay: free courses and certifications, in-person clubs, an agency partner programme, a yearly recap of what each user built, and executives posting on LinkedIn in their own words. A separate growth team turns that attention into pipeline with a livestream series built around use cases, a weekly list of accounts handed to sellers, and call blocks booked for the days after each episode. Clay says more than half of its pipeline now traces back to marketing and that there is no cold outbound in the process. Those are Clay's figures, from a post on its own blog in July 2026 by Fadeke Adegbuyi that wrote up a livestream by Bruno Estrella, who runs marketing, and Davide Grieco, who runs growth. If you want the primary source, watch the livestream and read the post. Every number in this piece comes from Clay unless I say otherwise.
How it started
Clay was founded in 2017 by Kareem Amin and Nicolae Rusan, and Varun Anand joined as a co-founder in 2021, per BetaKit's coverage of the funding rounds. The first product was not for salespeople. Clay's own account says the company spent about seven years going from zero to one and then about two years going from one to US$100 million in annual recurring revenue, a "nine-year overnight success" in the post's words. Amin told Rippling's founder blog in March 2025 what those seven years were like: "We went through loops, analyzing markets, building new features, and pivoting constantly." The narrowing came from watching who actually used the tool. Amin admitted he resisted it: "I had hangups about committing to being just a go-to-market tool."
The users who settled the question were B2B agencies running outbound for clients. By early 2023, per Mishti Sharma's account on First Round's site (she joined Clay that year), Clay was a spreadsheet-interface tool that outbound email agencies used to reach several databases in one place. Amin's team joined the agencies' WhatsApp groups to learn what they needed. "This group gave us some of our most sophisticated users," he told Rippling. Sharma's version of the distribution is the part to notice. Agencies and operators kept telling their audiences online how much time Clay was saving them, and an early playbook spread through popular sales WhatsApp groups. Nobody at Clay planned that. It was the behaviour the later machine was built to make repeatable.
The other founding act was a job title. Clay says it coined "GTM engineer" in 2023 (Sharma and Anand, writing on Clay's blog in April 2026): the person who sits between sales, operations and data and builds the plays. The same post counts about 100 job listings a month for the role. Estrella arrived from Webflow, where he had run growth, and by the July post's account made a bet on brand about two years before it was written. Grieco joined from Verkada, where he had been one of Clay's first customers in 2023, and started assembling a growth team about nine months before the post, with five people. Sharma's figures for her three years: ten employees to 400, US$1 million to US$100 million in ARR, a US$50 million valuation to US$5 billion.
How the machine works
The users make the content. Estrella's three convictions, per the post, are that people are the distribution lever, that "more is more" (the frequency of mentions across many places), and that users should feel loved. The programmes follow from that. Clay University is a free curriculum on AI skills, CRM enrichment and automated outbound, with certifications, and a certification is a thing a person posts about. Clay Clubs are in-person user meetups, with clubs started in London and Manila. The Solutions Partner programme routes client work to agencies and freelancers, the "Claygencies"; Clay says some have been acquired and some now bill more than US$1 million a year. Clayback is a yearly recap of what each user built, in the manner of Spotify Wrapped, and it exists to be shared. One flag for anyone reading the write-ups of this company: the Clay Creators programme they mention was, as of 19 December 2025 per Clay's own FAQ, "under construction and not open to applications". The machine is the users, not a formal creator roster.
The executives write their own posts, to a plan. Marketing runs an executive LinkedIn programme in which each target persona is paired with the matching Clay leader. A head of growth reads Clay's head of growth. A chief revenue officer reads a co-founder. Clay says the programme produced six million organic impressions in a recent quarter with no paid spend. Grieco's posting mix, as published: about 20 percent promotional with a clear call to action, 25 to 40 percent personal with none, about 40 percent thought leadership with none, and up to 15 percent playful. The executives write the posts themselves, sometimes with a teammate sharpening the hook, and the post says none of it is AI-generated. The pairing is what makes it a system. A buyer who follows three Clay people is not reading the same company caption three times.
The growth team makes content that has to pay. "How Clay Uses Clay" is a livestream series the growth team built, each episode on one use case rather than a feature tour. The first half is pitched at senior buyers; the second half is an operator building the use case live. Grieco's rule, which the post says he repeats about once a week, is that people care about use cases and pain points, not features. Clay's figures: around 19,000 registrants across the series, a blended cost per lead under US$12, and more than a third of sign-ups sourced through social. One of Grieco's own LinkedIn posts, promoting the episode about the marketing machine itself, drove more than 4,500 sign-ups. Grieco's line on the discomfort of promoting your own webinar: "Maybe it's cringe, but it's even more cringe having zero leads going to your webinar."
Half the job is selling it inside. Clay's growth team spends about half its time on internal marketing to its own sellers. Every Monday a "campaign in a box" lands: a ready-made list of accounts in Clay, with the invites a couple of clicks away. After each livestream episode, call blocks are booked for the following days so the leads are worked while they are warm. The post is blunt about why: "You can have the best idea, but if sales doesn't care about it, it's going to flop." It is also candid about what happened when the machine outran the sellers. Early on, with about a dozen reps, Clay disqualified 85 percent of demo requests because it had no capacity to take them. It tripled the rep count and built a team of real people, not AI, to handle the demos.
Around those four sit the slower parts. Repurposing: long videos are cut to 30-second spots for Instagram, past episodes are to be run as on-demand ads, and on the search side Clay's SEO lead Sung Jo wrote in April 2026 that webinar transcripts are turned into tutorial pages by Clay's own agent, that a library of more than 8,000 company and executive profile pages is refreshed automatically, and that an article now takes about three days instead of two weeks. "Every webinar Clay has ever recorded is a potential source of multiple template pages," that post says. Billboards ran in San Francisco and New York from September to November, and the way they were measured is the detail I would steal. The team scans Gong call recordings for billboard mentions, pulls them into a Clay table and links them to pipeline.
What it moves for the business
Clay's claims, labelled. More than half of pipeline traces back to marketing, with no cold outbound. Enterprise ARR roughly tripled in the nine months after the growth team formed. Sales-led revenue went from about 20 percent of the total to 33 percent, or 40 percent counting mid-market sales assist (the post's introduction rounds this to 40 and does not reconcile the two). The livestream series delivered about 19,000 registrants at under US$12 each. None of those figures has been checked by anyone outside Clay, and the post that carries them is marketing. What makes me take them seriously is the shape of the claims rather than the size. A company inflating its numbers does not usually publish that it threw away 85 percent of demo requests for lack of staff, or that it had to fix the middle of its funnel because brand was fine and capture was weak. The outside figures line up with the direction: a US$3.1 billion valuation at the Series C in August 2025 and US$7.1 billion at the Series D in September 2026, per BetaKit, with Clay saying ARR tripled to US$100 million in a year.
The honest limit of the evidence is attribution. Clay does not say how much of the pipeline comes from the user content versus the executive posts versus the livestreams, and I would not expect it to. What it does show is the pieces feeding each other. A certification becomes a post. A post becomes a webinar sign-up. A sign-up becomes a Monday list, and a Monday list becomes a call block. The measurement is pipeline, counted at the end, with Gong transcripts as the receipt.
What to steal
Make the users the publishers, and give them something to publish. A certification, a club, a yearly recap. Each one is an occasion for a post you did not write and did not pay for. Design the occasion, not the content.
Pair each buyer with one person who speaks to them. Not a company page. One leader per audience, writing in their own words, on a posting mix they can keep up for a year. The pairing is the system and the posts are the output.
Build one piece of content that has to pay for itself. Clay's livestream series exists because the growth team owns a cost per lead. Pick one series and give it a number it must beat. Everything else can be brand.
Spend half your content effort on the people who follow up. A Monday list, a call block the day after, an hour of training before a launch. If sales does not work the leads, the content did nothing, and the content team is the one who should fix that.
Measure with the receipts you already have. Call recordings are a record of what buyers mention. Scan them for your campaigns before you buy an attribution tool.