The Recommendations Guy
Leon Lin
He decided what hundreds of millions of people saw on Pinterest. Now he decides which creators a brand should actually bet on.
He decided what hundreds of millions of people saw on Pinterest. Now he decides which creators a brand should actually bet on.
Leon Lin runs 1stCollab, a San Francisco company that does an unglamorous thing extremely well: it takes the manual slog out of influencer marketing. Sourcing creators, sending outreach, negotiating, contracting, paying, even handling the taxes - the platform automates the lot, then measures whether any of it actually drove revenue. The pitch he and his co-founders make is blunt: do for creator campaigns what Google and Facebook did for digital ads.
The interesting part is who is making that pitch. Leon is not a marketer who learned to code. He is a machine-learning person who learned that the hardest question in influencer marketing - which creator is worth the money - is a prediction problem. And prediction problems are the thing he has spent his entire career on.
1stCollab is a Y Combinator company from the Winter 2023 batch. By 2024 it was reportedly running at around $2.1M in revenue with a team of roughly fourteen people - the kind of revenue-per-head ratio that tends to come from automation actually working, not slideware.
Something surprising about finance and tech to help you avoid boring people.
Leon joined Pinterest in 2013 as an engineer and stayed nine years - long enough to watch a scrappy pinboard turn into a place where hundreds of millions of people go looking for ideas. He moved from engineering into product, then up to leading Discovery: Home, Search, and the Related surfaces. In plain terms, when you opened Pinterest and it guessed what you wanted, his teams had built the machinery doing the guessing.
The work was about distribution - growing native content and video, lifting engagement, and increasingly reshaping the feed around creators rather than just pins. That last bit is the hinge of the whole story. Spend three years thinking hard about how creators get discovered and monetized, and you start noticing where the system leaks.
What he kept seeing: brands wanted to work with creators, creators wanted to work with brands, and the matchmaking in between was a mess of cold DMs and spreadsheets. A decade of ranking content for users, it turns out, is excellent training for ranking creators for brands.
Do for influencer marketing what Google and Facebook did for digital ads.
The founding trio knew each other from the inside. Leon was Head of Discovery. Varun Bansal had been a product manager on recommendations and creator engagement, with a math-and-CS background from Harvard. Andrew Liu had been a tech lead on search and recommendations at both Coupang and Pinterest. Read that lineup again: it is three people who built discovery systems for a living, deciding to build a discovery system for the creator economy.
The problem they picked is the one every growth marketer complains about. You send twenty cold messages and hear back from two. You spend a week negotiating, contracting, and wiring payment for a single creator. And when it is over, you genuinely do not know if it worked. 1stCollab attacks all three: instant sourcing across TikTok, Instagram, and YouTube; automated outreach, contracts, payments, and taxes; and real-time performance tracking that ties spend to revenue, complete with A/B testing.
Early brands on the platform include Framer, Locket, Gamma, and SnapCalorie - the kind of product-led startups that live or die on efficient growth. The company says it has saved clients more than $1M a year in influencer spend, the sort of number that comes from cutting the creators who look good but do not convert.
There is a tell in how Leon spends his time away from the cap table. He writes a newsletter called "Avoid Boring People," sending readers "something surprising about finance and tech" each month. More than 5,000 people subscribe. He has run it long enough to attach a book club to it - a quarterly commitment, meeting monthly, working through themes like the history of technology. The premise is quietly optimistic: being interesting is not a personality you are born with, it is a practice.
His own list of interests is short and a little wry - cocktails, art, and, in his words, being proven wrong. That last one is rare enough in a founder to be worth dwelling on. The whole 1stCollab bet rests on measuring whether campaigns worked, which means being willing to learn that your favorite creator was a dud. A person who enjoys being proven wrong is exactly who you want building that.
His X handle, @nilnoel, is "Leon Lin" spelled backwards.
It is the kind of detail that tells you more than a bio could: a person who finds the pattern, flips it, and ships it without commentary. The same instinct that ranks a feed will happily reverse a name for fun.
Make influencer marketing as measurable, automated, and obvious as buying an ad.