I thought I was building features. I was really building proof.

Day 11 on AffiSpark.

This week looked productive in the normal founder way.

I added promo-code attribution.

I added a public preview.

I replaced a fake walkthrough CTA with a real in-browser flow.

On the surface, that looks like feature work.

I do not think it was.

I think early product work is usually proof work.

Users are rarely waiting for your tenth feature.

They are waiting for proof that the risky part of your product is real.

For AffiSpark, the risky questions became obvious once I stopped pretending they were not:

Can attribution survive messy checkout behavior?

Can I inspect the product before paying?

If I click the CTA, can I actually finish the next step inside the browser?

Promo codes were proof that attribution does not fall apart the moment the sale leaves the neat referral-link path.

The public preview was proof that I am not asking for a blind purchase.

The walkthrough form was proof that the CTA can actually be completed, not just clicked.

Same week, same commits, different interpretation.

This matters because feature work sounds like momentum.

Proof work is what reduces buying anxiety.

A lot of founders build the roadmap as if users are asking:

what else does it do?

Usually they are asking:

why should I trust this part?

That question changes what gets built first.

If the core doubt is attribution, promo codes can matter more than prettier analytics.

If the core doubt is trust, a preview can matter more than another dashboard.

If the core doubt is completion, an in-browser flow can matter more than an extra settings page.

My current rule is simple:

find the sentence a user needs to believe before paying, then build the shortest thing that makes that sentence true.

Not persuasive.

True.

I think early-stage products get clearer when you stop measuring work by how many features shipped and start measuring it by how much uncertainty got removed.

Curious how others think about this:

what was the first piece of proof you had to build before your product started making sense?