Why you ship exactly one change

Attribution requires restraint

If you fix the page, publish two new articles, get a review site listing, and update your schema in the same week, and the number moves, you have learned that something worked. You have not learned what, so you cannot do it again on purpose.

One change per cycle is slower per week and much faster per lesson learned. After three cycles the team that changed one thing at a time knows three things. The team that changed fifteen things knows that their numbers went up.

This is genuinely hard to hold to under pressure to show activity. The defence is to be explicit that the cycle is an experiment with a stated hypothesis, not a sprint of content output.

Write the hypothesis before you ship

State it in one sentence: because the engine is failing at [mode], I am changing [thing], and I expect [specific prompt] to show [specific movement] by [date].

A hypothesis that cannot fail is not one. "This will improve our AI visibility" is unfalsifiable. "This should make us appear in the answer to prompt 4, which we currently lose to a named competitor" can be checked, and can be wrong, which is what makes it worth testing.

Write down the date you will re-measure, and make it at least two weeks out.

Where this lives in SolCrys

• The Actions queue holds promoted recommendations and tracks state from open through shipped.
• Keeping the action in the queue rather than in your head is what makes the later comparison possible — it timestamps the change.
• Record your hypothesis on the action so the next person to look knows what was being tested.

Lab: Choose and record one change

Needs a free SolCrys account (no credit card). No quota consumed.

1. Pick one action, matched to the failure mode you named in Module 5.
2. Write the hypothesis in the one-sentence form above.
3. Set the date you will re-measure, at least two weeks out.
4. Deliberately do not ship the other things on your list this cycle.

Start Free — free plan, no credit card.

Check yourself

• Can your hypothesis be wrong? If not, rewrite it until it can.
• Does the change you picked match the failure mode you diagnosed, or the one that is easiest to fix?
• What are you deliberately not doing this cycle?

Go deeper

This lesson is the spine. These guides are the depth.

Continue

• Previous: When presence is not the problem
• Next: Fixing a page so an engine can use it