Yesterday's lesson was: ask a smaller question earlier.

Today's lesson is: ask the right question for the right stage.

A generic feedback question creates generic feedback.

Pricing hesitation, setup friction, and activation failure are different problems.

Day 4 on AffiSpark.

Useful lesson from this week: onboarding problems are often sequencing problems.

The wrong screen shown too early can create more friction than a missing feature.

For paid SaaS, pre-payment and post-payment should not feel the same. If the user has not activated yet, they usually do not need the whole product. They need the next required action.

I used to think better feedback meant bigger forms and more open-ended questions.

I think the opposite now.

If a user is already leaving, any feedback? is too much work.

So I added anonymous 1-click exit feedback across the landing page, pricing, and product flow.

I think early-stage founders overrate roadmap depth and underrate first-session clarity.

The first 10 minutes matter more than the next 10 features.

If a user cannot reach one believable outcome quickly, the rest of the roadmap is downstream of a trust problem.

That has been changing how I prioritize product work.

This week I got useful feedback from comments.

But I think the more expensive objections are usually the ones nobody writes down.

They just leave.

A comment tells you what one person noticed.

This week looked like feature work: promo-code attribution, a public preview, and a real walkthrough flow.

I do not think it was feature work.

I think early product work is usually proof work.

Users are not asking "what else does it do?" first.

This week the most useful feedback did not sound encouraging. It sounded skeptical.

"Paying before seeing the setup is weird" led to a public preview.

"Request walkthrough doesn't work" led to a real in-browser request flow.

Both comments were more useful than praise because they named the exact place trust broke.

I got one of the best bug reports an early founder can get: "request walkthrough doesn't work."

The weird part: nothing was technically broken.

The CTA was just a mailto link.

That means it only worked if the user had a local email client configured and was willing to leave the product.

Today I separated two ideas I had been treating like the same thing: paid entry and blind entry.

I still want pricing visible early. That saves time and filters for real intent.

But asking people to pay before they can inspect the setup is just blind friction.

So I did not back away from paid entry. I added a public preview instead.

Today I fixed a mistake before it became product debt: I almost made affiliate links the only attribution model.

Links are clean, but they assume the click and checkout stay connected.

Real buying behavior is messier than that. Extensions, copied checkout URLs, cross-device purchases, and word of mouth all break link-only attribution.

In those cases, the link starts the journey, but the promo code is what survives to checkout.