The Kill/Scale Decision Tree

10 min · interactive · Advanced

Every operator has a product in "maybe" mode — it has ROAS of 1.7, conversion rate of 1.1%, CPA trending upward. Kill it or scale it? Here is the decision tree that removes the emotion and lets the data speak.

The decision framework

Do not gut-check. Do not pray. Do not "give it one more week." Use this tree.

Step 1: Is the 7-day ROAS above 1.5?

Yes → continue to Step 2. No → Kill it. Pause the campaign same day. Do not wait for more data. A sub-1.5 ROAS means the unit economics are broken. Beginner mistake is hoping it improves — it rarely does.

Step 2: Has the product been live for 14+ days?

Yes → continue to Step 3. No → Keep testing. You do not have enough data. Minimum 14 days / ~200 orders before making a kill/scale decision.

Rationale: Week one is pixel-learn phase. Meta's algorithm is still calibrating. Real ROAS emerges around day 7-10. By day 14, you have a reliable signal.

Step 3: Is the 7-day conversion rate above 0.8%?

Yes → Skip to Step 5 (Scale). No → Continue to Step 4.

Rationale: Above 0.8% means your product-audience-creative fit is strong. The machinery works. ROAS drift is usually audience fatigue (solvable). Below 0.8%, either the product doesn't match the audience, or the creative is not landing.

Step 4: Have you tested 3+ creative variants?

Yes → Kill it. The product does not convert at acceptable rates. Move to the next product. No → Rebuild creative and re-test. One more week with a new hook/format.

Rationale: If 3+ creatives all produce sub-0.8% conversion rate on the same audience, the problem is product-audience fit, not creative. Kill it.

Step 5: What is the trend on CPA (cost per acquisition) over the past 7 days vs. the 7 days before?

CPA is flat or declining → Scale. Double the budget. Duplicate the top-performing ad set. Go.

CPA is rising 10-20% → Keep testing. Add one new creative variant. Pause the oldest creative variant. Hold budget flat.

CPA is rising >20% → Rebuild creative (fatigue signal). Test a new hook, format, or angle. Hold budget flat until conversion rate stabilizes.

The scaling playbook (post-Step 5 decision)

Once you have decided to scale:

Duplicate the top-performing ad set. In Meta, duplicate the exact audience/placement/creative combo. Start at 0.5x current budget (e.g., if current is