UGC Handbook / Lesson 02 of 14 / Chapter I: Before you post

Decide what winning means before you post

"You only get to see the TikTok that does 10 million views. You don't get to see how many of those actually convert into paying users."

That is David Barnard on the Sub Club podcast, in an episode with Dan Layfield, who consults with startups on subscription revenue. He added the part nobody posts about: how many of those paying users stayed subscribed, and whether it was a good cohort at all. "The answer is often it isn't."

The 10M-view video is the only part of the funnel that is public. Everything that decides whether you made money is private. So before you post, you have to decide which private number you are actually playing for.

Views are an input. Revenue is the output.

Every UGC (user-generated content) video runs the same six-step funnel, whether you track it or not.

View → Click → Install → Trial → Paid → Revenue.

The first step is the one everyone can see, and it is also the one you can manufacture. One founder who took an AI app to $300K a month in 45 days, speaking in a Superwall interview, wrote views as a plain multiplication: number of creators × videos per creator × posts per video × views per post. Push any of the four up and views go up.

Nothing in that equation mentions a paying user.

The platform is not going to fix this for you. As one Superwall host put it, TikTok's organic algorithm is optimizing for views, not conversions. Unlike a paid ad, there is no conversion event to tell the feed "show this to more people like the ones who just paid," so it shows the video to more people regardless of who converts.

Views are an ego metric. The founder of the couples app Flame said exactly that in a Superwall interview: he wanted views early on, then realized "views don't matter if your videos don't convert." His filter is the comment section. If people are not asking "what is the app, where can I get this," it is content that gets views, not content that converts.

So treat the funnel as two halves. The left half (View, Click) tells you whether the video works as content. The right half (Install, Trial, Paid, Revenue) tells you whether it works as a business.

Only the right half pays the bills.

A video can win at views and lose at revenue

Here is the trap in numbers. The figures below are invented to show the pattern, not a benchmark.

Which one won? It depends entirely on the decision you are making.

• If the question is "which hook stops people," Video A wins. More than double the views.
• If the question is "which video brings users who want the product," Video B wins. Nearly twice the trials from fewer installs.
• If the question is "which video deserves the next batch of creator budget," Video B wins by a lot. Four times the paid subscribers.

Same two videos, three different winners. There is no "best video" until you say what it has to win.

This is the Barnard point in miniature. Video A is the one you would see on the feed. Video B is the one you would want to have made.

Why the down-funnel numbers move the other way. RevenueCat's blog put the cause plainly in a post on UGC ads: old UGC "optimized for one metric: the click." Content that grabs attention without setting expectations gets the click and then loses the user at install or at trial. Content that earns the view with a real problem and a real product gets a higher-quality click and more subscribers.

Annoying, but true: the video that looks worse on the platform can be the one that pays.

Pick one primary metric per test

You can watch every number. You should only let one of them decide.

The primary metric is the one that gets the final vote. Before the video goes live, write down what you are testing and which single number tells you whether it worked. Everything else is context.

Two rules make this table work.

Rule one: the metric has to match the decision. If you are deciding where next month's creator budget goes, judging by views is judging the wrong half of the funnel, so use paid subscribers. If you are deciding which of five hooks to keep filming, paid subscribers will take weeks to accumulate. Use view rate or clicks there, and accept that you are learning about content, not about revenue.

Rule two: add one guardrail. A guardrail is the number you refuse to make worse. A hook test can win on view rate and lose on trial starts because it attracted the wrong people. Write the guardrail next to the primary metric so the win has a condition attached.

Now fill this in. One line per test, written before you post:

The first blank is one thing (a hook, an angle, a creator). The second blank is one metric from the table. If you cannot fill both blanks with one item each, you are running two tests and will not be able to read either.

What about cost? Creator fees do not flow into Airbridge on their own. If you want cost per paid subscriber for an organic creator, it is arithmetic you do by hand: your creator spend ÷ paid subscribers from Airbridge. Cost-based metrics inside Airbridge appear once you connect an ad account, which matters when you boost a winner (Lesson 13).

Do this before you post

1. Write the funnel on paper. View → Click → Install → Trial → Paid → Revenue. Mark which steps you can see today. If everything right of Click is blank, Lesson 03 is your next stop before any video goes live.
2. Name the decision this test serves. "Which hook to keep filming" and "which creator gets next month's budget" need different metrics. Pick one decision per test.
3. Fill in the sentence. We are testing to see whether it increases . One item per blank.
4. Add the guardrail. The number this test is not allowed to hurt. For a hook test, usually trial starts. For a creator test, usually paid subscribers.
5. Set the decision rule now. Write what you will do if the primary metric wins, ties, or loses: Scale, Iterate, Watch, or Kill. Deciding after the numbers arrive is how every video becomes a winner at something.
6. Post, then judge only against the sentence. Views are allowed to be interesting. They are not allowed to vote.