SaaS Growth Vault
Playbook·Idea → Evidence

20 Conversations Before Code

Find evidence before you hide in the build.

STATUS
IN USE
SOURCE
MENTORING
UPDATED
22 JUL 2026
EVIDENCE
Radwan's operating rule for early validation.
6 min read

The problem

You have an idea. It feels obvious. So you start building.

Then you spend three months on a product no stranger has asked for. The problem was never the code. It was that you skipped the part where you find out if anyone actually has this problem.

Why founders get it wrong

Building feels like progress. Talking to people feels slow and scary. So we pick the comfortable work and call it momentum.

Failure mode

You show the idea to friends, they are nice about it, and you treat "that's cool" as demand. Praise is not evidence.

The decision rule

The decision rule

Do not write production code until you have had 20 real conversations with the narrow buyer, and at least a few have described the problem without you leading them.

The plan

  1. Name one narrow buyer you can actually reach.
  2. Book 20 short conversations. Ask about their problem, not your solution.
  3. Listen for the problem in their words. Note who describes it unprompted.
  4. Ask for one small commitment: money, time, data, an intro, or a signed intent.
  5. Decide: continue, change the buyer, or stop.

Example

Interview opener

Hi — I'm looking into how small SaaS teams handle [problem]. Not selling anything. Could I ask you three questions about how you do it today? 15 minutes, your call on timing.

Do this today

Do this today

Message five people who might have this problem. Ask for 15 minutes. No pitch.

What good looks like

People describe the pain before you do. A few offer something real to keep talking. You can quote the problem back in one plain sentence.

Checklist

Checklist

  • One narrow buyer named
  • 20 conversations booked
  • Asked about the problem, not the solution
  • Tracked who described the problem unprompted
  • Asked for one real commitment each time

Evidence and limits

Counterexample

Enterprise products may need fewer but deeper conversations. Low-price consumer products may need more behavioural data than talk. Twenty is a floor for a typical B2B SaaS idea, not a magic number.

Next page

Run the Idea Evidence Score with what you learn.

Evidence and limits

A practical heuristic, not a statistical law.