When to say no to an AWS program

Checking whether you qualify for an AWS partner program takes minutes. Satisfying one can take a sprint. Almost everything that goes wrong with partner programs lives in that gap, because the cheap half of the decision is the half that gets automated and the expensive half is the one nobody prices.

Which makes a list of every program you're eligible for a worse artifact than it looks. AWS runs more than eighty of them, and a functioning partner qualifies for far more than it should pursue. "Eligible" and "worth it" are different questions, and only one of them has an API.

Eligibility is a fact; fit is a judgment

In our product we support exactly four programs with enriched eligibility detail, chosen by expected value, and that ratio (four of eighty-plus) is not a coverage gap we're racing to close. It's the same discipline we're recommending here, encoded: the check must be cheap, but the yes must be expensive.

What we do, for ourselves and in how the product presents findings, is price the yes before saying it. Three questions. What is the benefit worth in dollars or in pipeline, honestly discounted by the probability it materializes? What does compliance actually cost, counted in the hours of the specific people who'd do the work, who are never idle people? And the one that catches most programs: does the work produce anything you'd want even if the program vanished?

That third question does most of the filtering. Work shaped to satisfy a checklist tends to be worth very little once the checklist is gone, and enablement artifacts in a required format are the clearest case: real effort, real deadlines, and no reader outside the program. Migration and POC funding on live deals sit at the other end and pass trivially, because there the work is the deal.

There's one legitimate exception. Occasionally a program's real payoff is standing with an AWS team you need, and the artifacts are the tuition. Fine, but then say that sentence out loud in the meeting ("we're doing this for the relationship, expecting no direct return") and watch how many programs survive it. Vague strategic hope is how long detours get approved.

Real numbers

1. Minutes to check, weeks to satisfy: the asymmetry that makes the yes the expensive half of the decision.
2. 80+ programs in the wild; 4 we support with full eligibility detail, on purpose.
3. 3 questions per program: honest expected value, true compliance cost, would-we-want-the-work-anyway.
4. 1 exception: relationship tuition, but only when it's said out loud.

Where the humans sit

An eligibility engine can say "you qualify" at scale; only a human can say "and it's still not worth it", because the answer prices your team's next sprint against your roadmap, and no program's terms contain your roadmap. In our product the findings arrive as recommendations with reasoning, and declining one is a first-class action, not a failure state. Some of our proudest eligibility results are the ones a customer read carefully and closed.

Steal this

List every AWS program you're currently in or applying for, and run the three questions on each, with the compliance cost in named people's hours, not abstract effort. Drop anything that fails all three and isn't consciously relationship tuition. Then cap the portfolio: pick the two or three programs your deal profile hits most often and go deep, because a program you're deep in (known to its AWS team, fast through its process) pays a repeat dividend that a portfolio of shallow enrollments never does. Saying no to a program isn't leaving money on the table. It's deciding what your team's next six weeks are for.

Next on this track: the map of what AWS actually funds, the human companion to the funding automation the product track takes up first.

---

This post is part of the partner track of How a Tiny Team Ships Like a Big One, a series on how six builders run a production AI company. Building at Aithon - if this is how you want to work, talk to us.