Why I paid creators for the video, not the views

Almost every guide to paying ambassadors lands in the same place: commission is the most ROI-friendly model, because you only pay for results. It sounds unarguable. You pay for outcomes, the creator carries the risk, nobody wastes money on content that goes nowhere.

I think it's wrong for developer tools, and I built the opposite.

There are two obvious ways to pay a creator, and both of them quietly ruin the content.

Pay per view and you're paying for reach you already priced into their tier. Creators respond by chasing titles and thumbnails, because that's what the money is asking for. You end up paying most for the content that teaches least.

Pay per signup or commission and you invite conversion spam. The fastest route to a signup is rarely a careful tutorial.

So I did neither. Flat fee per piece of content, set by tier, agreed before anyone starts.

Why "only pay for results" backfires here

The commission argument assumes the creator controls the result. For a consumer product bought on impulse, roughly true. For a developer tool, not close.

A developer watches a tutorial, doesn't sign up, and comes back nine weeks later when they hit the problem it solves. Or they sign up on a work email through a link nobody tracked. Or the tutorial does its job and the signup dies on your pricing page. None of that is the creator's doing, and all of it lands on their invoice.

So good creators price in the risk and ask for more, or pass. The ones who accept are the ones who most need the money — not the same people as the ones who explain your API well.

There's a second cost. Commission makes you a variable expense in someone's month, and anyone with a full pipeline schedules guaranteed work first. You end up last in the queue while believing you built an efficient program.

Pay for the thing you actually want

The point of a creator program isn't views this week. It's content that answers a real technical question, ranks on search, and keeps working a year later.

You get that by paying a known amount for defined work, then getting out of the way. A flat fee also makes income predictable, which is the difference between someone who fits you into their schedule and someone who plans around you.

Put the upside somewhere it can't distort the content — for us, a revenue share on referrals, sitting on top of everything else.

Three content types, defined tightly

A flat fee only works if both sides agree what was bought. That means writing down the content types before you sign anyone.

We used three.

Dedicated. The product is the subject. It's in the title, it's the reason the piece exists, and it holds most of the screen time. "How to scrape any website with X."

Integration. The product does a real job inside a larger build. It isn't the subject — the project is — but it gets meaningful screen time doing actual work. "I built a RAG app with X, Y and Z", where your tool handles one real step.

Mention. A credit inside something they were making anyway. Short, and it has to show something — the API call, the output, the node in the workflow. Saying the name without showing anything doesn't count.

That last rule matters more than it looks. Mentions are the format that performs best, because they don't feel like ads. They're also the easiest to fake. Requiring a few seconds of something real on screen is what keeps the format honest.

Points on top, for staying

The flat fee pays for a piece of work. It does nothing to reward someone for still being here in month six.

So points sit on top. Every piece of content is worth a fixed number based on depth — a dedicated video is worth twice an integration, and a passing mention is worth zero. Points add up for the life of the partnership and unlock bonuses at set totals.

Two details make it work.

Bonuses scale with the creator's own rate, not a flat cash amount. A bonus that feels meaningful to a small channel is a rounding error to a large one. Expressing it as a share of what they already earn keeps it meaningful at every size.

The ladder doesn't end. After the top milestone, the bonus repeats on a fixed interval, forever. A ladder with a final rung tells your best creators that the program is done rewarding them, right at the point they've become most valuable.

Why mentions earn nothing

This is the rule people push back on hardest, and it's the one I'd keep.

Mentions get paid — they're real work and they perform well. They just don't move you up the ladder. The ladder is for commitment, and a mention inside something you were making anyway isn't a commitment. If mentions earned points, the cheapest path to every bonus would be a stream of one-liners, and the deep content the program exists for would stop getting made.

Pay for everything. Only count what you want more of.

What I got wrong

The brief. The first cohort turned in weaker drafts because I hadn't been specific enough about what a good integration looked like. We fixed it over three rounds of feedback. That cost two weeks I could have saved by writing it properly upfront.

The brief is the most important artefact in the program. It's the difference between a creator who makes a tutorial worth watching and one who makes a feature mention nobody clicks. Write it before you recruit anyone.

Paying well gets people in. It doesn't keep them — that's a separate design problem, and the usual answer to it is backwards. I wrote about that in the streak system that never punishes creators.