Hector Serrano on Paid UGC: How Wispr Flow Runs 500 Creator Ads a Month

Hector Serrano, head of content at Wispr Flow, explains how a 50-creator paid UGC program produces 500 to 600 ads a month for the AI dictation startup. Serrano joined Wispr Flow at 19, three days after quitting Division 1 soccer. He spent a year and a half making every ad himself, then turned what he had learned into a roster of creators who are paid a cut of ad spend, with the best earning more than $13,000 a month. Over the past year Wispr Flow’s revenue grew 30x and the company raised at a valuation of about $2 billion. In this 87-minute conversation on Lenny’s Podcast, Serrano and host Lenny Rachitsky go through the full playbook, from finding the first five creators to why the best marketers are “allergic to AI” and why YouTube packaging matters more than production quality.

TLDW

Paid UGC means recruiting real users to make video ads, then paying them a share of what each ad spends (for software) or sells (for e-commerce), with a “smart retainer” as a floor while they ramp up. Serrano’s playbook: recruit a small, varied group by hand on Instagram, build a real community around it, give creators no briefs, and push them to experiment so Meta’s Andromeda system doesn’t group their ads together and wear them out. He argues that young, scrappy marketers with a following beat polished resumes. AI should stay away from anything the customer sees. Founder-led content and well-packaged YouTube make every paid conversion cheaper.

Thoughts

Meta’s Andromeda system is the real reason this program exists, and Serrano says so almost in passing. When an ad platform scans every frame, the transcript and the on-screen text, and groups near-identical ads together, the creative problem stops being volume and becomes variety. His numbers make the point. In 2024 an ad could run for twelve months. Now one dies at around $100,000 in spend, which can take as little as two weeks. A single talented in-house creator, which is what Serrano was for a year and a half, cannot keep up with that. Fifty creators from fifty different lives can, but only if they stay different. That is why he refuses to brief them, and why he keeps steering them away from copying whatever tops the spend leaderboard. The roster works less like a marketing team than like a defense against an algorithm that remembers everything it has already shown.

The “allergic to AI” line is more careful than the tweet made it sound. Serrano is not against automation. He names Matt Swulinski, Wispr’s former head of growth, as the most automated marketer he knows. His rule is narrower: AI can run the pipeline, but not the last 10% that the customer actually sees. Coming from an AI company, that is brand protection as much as taste. A dictation tool pitched as helping people sound more like themselves cannot afford ads that sound like nobody. The wider lesson is that as generated content fills every feed, visible human effort becomes the scarce signal. It is also one of the few signals a small team can supply cheaply.

The most underrated exchange comes near the end. Serrano says organic touchpoints make paid conversions cheaper, and Rachitsky adds that the paid platform then takes the credit. That is a quiet warning about how teams measure this kind of work. The founder’s YouTube episodes, Instagram reels and LinkedIn posts build the trust, and the Meta ad gets the conversion in the dashboard. A company that judges each channel on last-click numbers would cut the founder channel first, then watch its ad costs rise without knowing why. Wispr treats founder content as a long-horizon bet, with a soft sell perhaps a year in. That patience is part of why the paid machine looks as efficient as it does.

The pay design is clever and worth copying. A smart retainer (for example, $1,500 for 25 videos) only tops a creator up to that amount when commissions fall short. It carries new creators through the months before their ads find spend, then disappears once they earn more on commission. Paying a share of ad spend rather than of attributed sales avoids the attribution fights that freemium software invites. Meta only keeps spending behind ads that work, so spend is a reasonable stand-in for performance. The model has a soft spot, though. A creator’s income depends on the media buyer’s budget decisions, so the whole program runs on trust between the creative lead and the people he recruits. Seen that way, Serrano’s emphasis on weekly payouts and treating creators as friends is what holds the model together, not just culture talk.

Read the hiring advice skeptically. It is not wrong, but it comes from a survivor. A 2.5 GPA athlete who becomes head of content at 20 proves that resumes miss talent. It does not prove that any hungry teenager will work out. The useful part of the sports-academy comparison is the part Serrano dwells on. Wispr gave him and Tobin Tang a budget, months of visible failure, and patience through the “hard conversations.” Companies that copy the hiring profile without copying the patience will pay the cost of an academy and never see the graduates.

Key Takeaways

Product-market fit comes first. Serrano says there is no growth hack for a product that isn’t growing on its own, and paid, organic and creator channels only add fuel to a fire that is already burning.
Turn on paid early. Following an e-commerce approach he learned from Matt Swulinski, Serrano argues that paid ads tell you fastest whether your messaging can turn cold traffic into users. Most companies are not skilled enough at organic content to get a clean signal there.
Wispr’s first creator program was organic UGC paid per thousand views. It reached about 500 million views in two and a half months but drove few conversions, so the company killed it and moved the community-and-performance model over to paid ads.
Creators are paid a percentage of ad spend for software and a percentage of sales (GMV) for e-commerce. Spend-based pay suits freemium products where attribution is slow or unclear.
A “smart retainer” sets a floor, such as $1,500 for hitting a 25-video quota. Wispr only pays the gap between that floor and the creator’s commissions, and creators who out-earn it keep pure commission.
No briefs. Long guideline documents strip out the creativity that makes creators valuable, and some of Wispr’s best ads, like raindrops “typing” on a keyboard left out on a rooftop, came from creators’ own ideas.
Start with about five creators, found through manual, targeted Instagram outreach: 150 to 250 DMs a week, searching niche keywords and mining good creators’ follower lists. Mass-messaging software brings back low-quality creators.
Recruit for variety, not headcount. Each creator should speak to a different customer type (Wispr’s roster includes moms, lawyers, engineers, sales reps and an eight-figure e-commerce founder), because a hundred similar creators make a hundred similar ads.
Culture keeps creators: weekly meetings, creator coaches, giveaways, a retreat, payouts every seven days, and a $100 referral bonus for bringing in creators who get accepted. Wispr caps the group at 50 and now receives 1,000 to 1,500 applications a week through Trybe.
The team can be two people: a creative lead who recruits creators and approves videos, and a media buyer who runs the ad accounts on Meta, LinkedIn and YouTube.
For technical and older audiences, ugly ads win. Plain screen recordings with text on top are among Wispr’s best performers, and an expensive product should be pitched by someone who looks like its buyer.
Hire marketers who have grown an account to 10,000 or 15,000 followers, ideally on Instagram or YouTube. Serrano says job posts with high salaries and resume requirements attract the wrong people.
Keep AI out of the final creative. Automate the research and the workflow, but keep a human on the part customers see.
On YouTube, the title, thumbnail and story matter more than production value. A strong long-form edit also gives short-form editors better raw material. Serrano’s most recommended book is The Almanack of Naval Ravikant, for its idea of media as leverage that needs no one’s permission.

Chapters

08:27 From Organic UGC to Performance-Based Paid UGC

Serrano traces the program back to Tobin Tang, an early marketing intern who built an organic UGC group. Creators posted Wispr videos on their own side accounts and were paid per thousand views. It hit roughly 500 million views in two and a half months, but it chased vanity metrics and students rarely paid. Seeing that his own paid ads were working, Serrano applied Tang’s community model to paid ads instead. He then found Trybe, a platform built by SquidHaus founder Lukas Pakter that brings TikTok Shop-style performance payouts to Meta ads, which software companies could not use before.

13:44 How Creator Pay Works: Smart Retainers and Percent of Spend

Pure commission demoralizes new creators, so Wispr pairs it with a smart retainer. Hit the video quota and earn $500 in commission, and Wispr adds $1,000 to reach the $1,500 floor. Earn $3,000 and you keep commission only. Pay is uncapped, which is why some creators with 1,000 or 2,000 followers make more than full-time influencers. Because creators work with the same brand every month, their ads compound in quality, and new features can go to the whole group through Discord instead of a scramble for new talent.

20:07 The Profile of a Modern Marketing Hire

Serrano contrasts his resume with Tang’s: a 2.5 GPA and Division 1 sports against a 4.0 from Waterloo. Both turned out to be strong marketers. His bar is simple: a hard worker, a good person who fits the culture, and someone who has grown an account to around 10,000 followers. He rates Instagram and YouTube as better signals than TikTok, and notes he grew from 500 to 18,000 followers on X in two weeks once he learned it was a short-form volume game. He also argues that young people who live in Discord chats and Instagram feeds understand new channels better than any 40-year-old founder.

31:03 Building a Paid UGC Group From Zero

Set up the payout model, open a group on a platform like Trybe, then go off-platform to recruit, because creators won’t join a brand that has paid out $0. Serrano sent 150 to 250 targeted Instagram DMs over a week to land five or six good creators, then grew by five or six every few weeks. The danger at scale is creators copying the top-spending ads. Andromeda groups similar ads and tires them out quickly, so he keeps pushing creators to experiment. He adds that UGC for older technical buyers should come from people who look like them, and that ugly screen recordings often win.

43:03 Outreach, Onboarding and Creator Culture

The easiest first recruits already follow the brand or sit in its DM requests. Serrano’s opener is short and does not lead with the performance model: he liked their content and wants them in the creator group. On the onboarding call he gets to know the creator first, then walks through his background, the media buyer, the product, how Trybe works and the seven-day payout. He closes on the call so they can start filming right away. He says the companies that treat creators best will win long-term, because word travels fast.

48:31 Team Structure and the Media Buyer

To run paid UGC you need a creative lead (Serrano) and a media buyer (Eric Zaworski) who lives in Meta Ads Manager. Wispr has grown from about four people when Serrano joined to more than 120. The marketing team is about 15, flat, with no middle managers, all reporting to CMO Daniel McCallum. With more spend, ads tire out faster. Wispr has run about 5,000 ads in total, with 800 to 900 live, alongside statics, celebrity and influencer videos, launch films and founder content.

53:44 Why UGC Works Even in Technical and Enterprise SaaS

Enterprises are made of individuals, Serrano argues. Win one user with an ad, let them spread the product inside the company, then close the enterprise deal. Rachitsky notes this fits bottom-up, product-led tools better than a Workday or Salesforce. Serrano adds that creators with no following often beat influencers, because audiences now read big names as sellouts. For would-be creators, his advice is to stop asking for advice and start filming, since even famous podcasters stutter, refilm and swear at their cameras.

1:03:52 Why the Best Marketers Are Allergic to AI

Even top AI labs use large human teams for their final videos and designs, Serrano points out. An AI company that says it helps people sound like themselves cannot run AI slop ads without a backlash. Automate research and workflows, but keep a human on the last 10% the customer sees. The noise from mass-produced content makes quality easier to stand out with. Wispr founder Tanay Kothari’s YouTube channel reached millions of views and about 21,000 subscribers in five videos by posting rarely and with purpose.

1:08:07 Founder-Led Content as a Long-Term Play

People buy from people, and a founder on camera makes a product feel human. Wispr ties every episode to its mission of one billion daily active users, like a series where each episode differs but the theme holds. Episodes range from launching the note taker to Kothari’s wedding shopping in India. Serrano says the content can start before a brand is known, and that every organic touchpoint lowers the cost of paid conversions. His founder-content team is a head of content, a videographer, a YouTube packaging team and short-form editors, many of them hired through his own Instagram stories.

1:16:01 Why YouTube Packaging Matters More Than Production Quality

A Hollywood-quality video with a weak title and thumbnail goes unseen. Packaging covers the title, the thumbnail, the story, the pacing, the intro and a payoff that matches the thumbnail. A strong long-form edit gives overseas short-form editors a clean source to clip from. YouTube also has the longest half-life of any platform: one of Serrano’s personal videos went from 40,000 to 600,000 views in a year. He closes by urging founders to bet on young talent, and cites Steve Jobs on connecting the dots only looking backwards.

Notable Quotes

“When you send the creator a 10-page guideline on how they need to shoot their content, what they need to talk about, you end up removing the creative aspect that creators have.”
Hector Serrano, on why Wispr Flow gives creators no briefs

“That’s why you can reach ad fatigue really quickly if you’re constantly repeating the same formats, the same messaging, the same people.”
Hector Serrano, on how Meta’s Andromeda system treats look-alike ads

“The more boring and the uglier the ads are, the better they will perform and that can be something as simple as like a screen recording of someone using your product.”
Hector Serrano, on ads for older, technical buyers

“An ad once it hits like $100,000 in ad spend, will be dead and that can be hit in like a month or 2 weeks, 3 weeks.”
Hector Serrano, on why paid UGC needs constant creative volume

“I almost think that creators with no following end up being better investments for brands because influencers kind of are seen now as sellouts a lot of the time.”
Hector Serrano, on everyday creators versus influencers

“If you just think that throwing money out there and looking for someone with the best grades or the best school or the best experience is going to get you the best hire to run this for you, you’re in for a rude awakening.”
Hector Serrano, on the most common hiring mistake

“You can make a Hollywood production movie. But if you put a bad thumbnail and a bad title that doesn’t intrigue people to click on it, no one’s actually going to know how good your video is.”
Hector Serrano, on YouTube packaging

Watch the full conversation here.