UGC ad production

The operating routine for the committed UGC-ad roster, run by Product & Growth's UGC creator program (the Media Buyer), with the Creator in Residence contributing footage. paid-and-ugc-ads is the strategy: why Intempt runs paid media, why creator UGC makes the best-performing ads, and the terms that govern it. This page is the routine that runs underneath it, with owners, tools, and timing. creator-operations is the sister routine for sponsored placements and the Collective tracks. Where the two overlap on money and attribution, this page points to it rather than repeating it.

The spine: one recipe, one ad

Intempt's creative engine is its library of about 300 recipes, the named GTM plays across the five products: Data, Design, Marketing, Sales, Analytics. The rule that makes this program work is simple: one recipe equals one UGC ad. A creator picks a recipe, runs that GTM-engineering workflow for real on Intempt, films it in their own voice, and that footage becomes the ad. The ad shows the self-driving workflow people actually want to copy, not a feature tour.

Three things follow from that, and every step below leans on them:

• A near-endless supply of distinct creative. About 300 recipes, across 5 products, across the ICP personas, is a very large number of genuinely different concepts. That is the direct counter to Meta's similarity throttle (see the creative loop).
• Every ad is a runnable play. The recipe is publishable, so the audience can go run the exact thing they just watched in minutes on the free plan.
• The creative doubles as product. A filmed recipe is also a recipe other teams run and a referral surface, so one video can earn three ways at once (see the stacking section).

Who does what

Step 1: The recruiting funnel

Timing: a few days of concentrated outreach seeds each cohort. Owner: the partner manager.

Start small, on purpose. The first cohort is 4 to 5 creators, and each one is hired to fill a named gap, never "sign twenty and see." A gap is one of:

• A product lane with no coverage: Data, Design, Marketing, Sales, or Analytics.
• An uncovered persona: founder, SDR, lifecycle, analyst, or designer.
• A region or language the roster does not reach yet.

Source order. Work it warmest first and stop at the tier that fills the cohort.

Plan on roughly 100 to 200 DMs over a few days to seed the first cohort. The funnel is wide on purpose because the accept bar is narrow.

The accept bar is brutal, and it is measured, not asserted. The reference model (Wispr Flow) accepted about 5 to 7 creators from roughly 7,000 who came through the funnel. Intempt's posture is the same shape: a high bar, on the order of one in hundreds. Record the funnel so the real ratio can be read later, with a source and a date. Never state an accept ratio you have not counted.

Selection criteria, in order:

• Storytelling ability. Can they hold attention and land a point.
• Audience fit. The ICP personas, not guru or career-prep content. Reject the anti-personas in who-we-build-for.
• Production quality. Clean audio, watchable framing, a real editing hand.
• Genuine product understanding. They can run a recipe and explain it.
• Users-first preferred. Someone who already uses Intempt is a shorter path to a real ad.

Referral bonus. $100 for a referred creator who joins and is at or above the referrer's level. Beginners are excluded, so the incentive is proven creators pulling in proven creators.

Step 2: From recipe to ad

Owner: the partner manager, with the performance marketer. Timing: the brief is set when a creator joins and refreshed each month against a quota.

1. The performance marketer names the lane. From the ad account, the marketer says which product lane, persona, or problem needs fresh creative next. That is a demand signal, not a guess.
2. The partner manager turns a recipe into a brief. Pick a recipe that fills the lane, then write a short brief: the recipe, the persona it speaks to, the problem it solves on screen, the hook direction, and the facts the creator must get right. Give the frame, not a script. The nomenclature is Intempt and Blu Agent, never GrowthOS or Intempt AI, and nothing handed to a creator contains an em-dash.
3. Set the monthly quota. The order sets a video count for the month (reference: 25 videos a month maps to the reference retainer floor). Quota is how volume stays predictable.
4. The creator films the real workflow. They run the recipe on their own Intempt project, in their own voice, and record the real thing. A filmed recipe the audience can go run beats a staged demo every time.

Step 3: The weekly cadence

Owner: the partner manager and the performance marketer together. The weekly call is the heartbeat of the roster.

Run one call a week with the whole roster and the performance marketer:

• Walk the ad account. The buyer shows what is winning and why, live, on the real account.
• Show the leaderboard. Every creator sees where their ads rank. Radical transparency is the point, not a reward you dole out.
• Name what gets spend. The buyer says which creative is getting budget and which is being cut, so creators learn what the market rewards.
• Pick the next recipes and experiments. Decide together what to film next and what to test, tied to the gaps the account still has.

Between calls, coaches stay in the DMs and the partner manager keeps an open calendar for strategy time.

Step 4: The creative loop and the 60% rule

Volume comes from each creator making many distinct recipe-ads, not from many creators piling onto one winner. Roughly 50 creators producing this way yields 500 to 700 net-new ads a month.

The 60% rule. Meta suppresses reach for creative that is about 60% or more similar to what is already running (an operator estimate, not a published Meta rule). So the roster biases hard to net-new. The recipe library is the unfair advantage here: with about 300 recipes across five products and the personas, running out of genuinely distinct concepts is not a real risk.

A new hook is not a new concept. A new recipe, a new persona, a new problem, or a new demonstration is a new concept. A fresh caption over the same play is not.

The loop when a creator ruts, run by the coach:

When a creator copies the winner, the fix is always to point them at the experiment that produced it, so they run the process instead of cloning the output.

Scale

The roster grows in stages over about four months, and retention is the metric, not headcount.

One performance marketer runs the whole account at steady state (about 50 creators and 500 to 700 ads a month), because payout, whitelisting, and attribution are systematized. The metric that matters is retention: a creator still submitting at 90 days. Churn is the cost, not recruiting.

Culture

Run the roster like a coached sales team. A small group built on purpose outworks a thousand loose signups.

• Coaches. Creator coaches run the calls and live in the DMs. The partner manager and founder keep time open for strategy.
• Giveaways. Cash plus gear: cameras, mics, Meta Ray-Ban glasses. Reward the producers.
• The SF retreat. Top performers fly to the SF office for a week of coaching, product time, and brainstorming. Status and belonging are the retention mechanism as much as the money.

Measurement and pay ops

Measurement. The ad account is the source of truth for what the fee is paid on. For each creator, track the ad spend Intempt ran behind their videos (Meta attribution), the leaderboard rank, and the recipe-run and referral attribution on the Intempt side. Put a source and a pull date beside every figure, and never carry a number forward without re-pulling it.

Pay ops. The money flows and the dogfood attribution are run exactly as in creator-operations, Step 6 (invoice and pay) and Step 7 (measure). Two points specific to this program:

• The performance fee is paid on ad spend, net 30. The rate is 7% of the ad spend Intempt runs behind the creator's videos (set per order against CAC). It accrues while Intempt spends, it is uncapped, and it stops when the ads stop. Nobody is paid for a video that never gets spend. The base the 7% is computed on is Meta's reported ad spend for that creator's content.
• The smart retainer is a floor, not a fee. An order may set a monthly minimum tied to the video quota. Hit the quota: if the fee is below the floor, Intempt tops up to the floor; if the fee meets or beats the floor, there is no top-up and the creator keeps the full fee. It is time-limited by the order and exists to de-risk the first month or two. The worked example lives in creator-onboarding.

The dogfood point carries over: Intempt runs this program on Intempt. The creator's ?via=<handle> is stamped on the customer record at signup and recipe runs are attributed in Intempt's own analytics, on top of the Meta attribution that the fee is paid against.

Three ways at once

The UGC fee is separate from and additional to the two Collective tracks. One filmed recipe can earn all three, and taking the UGC fee never reduces the other two.

A creator who films a recipe, carries their ?via link in the post, and publishes that recipe is on all three at once. The UGC fee is agreed per order; the Share and Build tracks are self-serve in the creator's Intempt profile.

A note on the numbers

Tune the spend intensity to Intempt's CAC and goal. Every real figure in a live order or report carries a source and a pull date.

Related: paid-and-ugc-ads · creator-operations · creator-onboarding · creator-in-residence · media-buyer · who-we-build-for.