Four common stages: test, rank, scale, harvest
Without stable conversion history, do not launch with a mature listing’s exact structure and low ACOS target. Move by stage, with one primary goal per stage.
- Testing: limited budget to find relevant queries and obvious junk. The goal is sample size, not the lowest ACOS.
- Rank building: give verified relevant terms more exposure when stock and reviews allow. The goal is organic movement; a higher ACOS can be temporary.
- Scaling: concentrate budget on converting terms and stable targets. Raise budgets there instead of expanding every match type at once.
- Harvest: bring target ACOS back below break-even, grow the winners, and stop paying a rank premium.
Gates for leaving a stage
Do not advance on a calendar. Use inventory, reviews, whether relevant terms convert, and whether organic rank is moving.
- Leave testing when you have a set of relevant search terms, junk is negated or isolated, and budget is not still buying obvious irrelevant stems.
- Enter rank building when sellable stock can cover faster orders, the offer is not obviously broken, and reviews have at least started.
- Enter scaling when several core terms still convert above your click threshold, ACOS is explainable, and organic rank is not falling.
- Enter harvest when organic traffic can take some demand and paying a high ACOS for rank has weaker marginal effect.
- In any stage, if stock drops under about 14 days, quality complaints appear, or the listing changes materially, return to watch or pause scaling.
How budget, match type, and ACOS should change
- Testing: small budget, auto or broad/phrase observation, ACOS may sit above break-even, with a hard daily cap.
- Rank building: shift budget to verified terms; exact and strong phrase first. Account ACOS may sit near break-even with an end date.
- Scaling: raise budgets on working ad groups, not equally across every campaign. Start pressing target ACOS down.
- Harvest: target ACOS clearly below break-even. High-ACOS leftovers get lower bids, tighter match, or a pause.
Example: the same SKU should not keep one target for four weeks
Break-even ACOS is 32%. Weeks 1–2: $20/day testing, some terms at 45% ACOS, goal to find 8–10 relevant terms and negate junk. Weeks 3–4: 180 units in stock, 20 reviews, budget concentrated on 5 converting terms, account ACOS allowed to 35% while watching organic rank. If rank has not moved by the end of week 4 and reviews are still thin, do not add more rank budget — fix the listing and review gap.
Risk boundaries: where mixed stages lose money
- A mature-listing low ACOS target in testing closes relevant terms before they convert.
- Rank pushes without stock can sell out, then organic and ads fall together.
- Scaling every match type at once also scales junk queries.
- Do not treat a fixed CTR or CVR as an official gate to the next stage. Use your conversion cost and rank change.
Use reports to see which stage you are actually in
If the Search Term report is still dominated by zero-order spend and irrelevant queries, you are still testing. If winners are stable but budget is still spread across broad match, you should be scaling or harvesting. Upload the file and read it against the stage goal, not against one ACOS for the whole year.