The Auto Campaign That Looks Great but Isn't
An advertiser runs a single Sponsored Products auto-targeting campaign for a $60 product. It reports a 19% ACOS over 30 days on $1,900 spend and $10,000 sales - better than the account's 28% target. The advertiser plans to scale it by raising the budget 3x and adding a high Top-of-Search placement modifier. Before scaling, a search term report shows: 62% of clicks come from close-match, 21% from loose-match, 9% from substitutes, 5% from complements, and 3% from close-match of the brand's own name. Loose-match converts at 4% vs. close-match at 14%. The product has a 12% return rate, and returns are not reflected in the reported ACOS.
1. Should the advertiser scale this campaign as planned? Identify the hidden problems in the reported metrics.
Read the worked answer
No - do not scale as planned. The 19% ACOS is real but flattered by two things the report hides.
Returns are invisible. A 12% return rate means ~12% of the $10,000 in attributed sales will be refunded. Adjusted sales ≈ $8,800, so true ACOS ≈ $1,900 / $8,800 ≈ 21.6%, not 19%. Still under target, but the margin for error is thinner than it looks.
Loose-match is a drag disguised by close-match. Loose-match is 21% of clicks but converts at 4% vs. 14% for close-match. That segment is almost certainly unprofitable at a $60 price point. Scaling 3x multiplies that waste 3x.
The auto campaign is a discovery engine, not a scaling vehicle. Auto-targeting's job is to surface converting search terms. The right move is to harvest the winners into a manual exact-match campaign and let auto keep discovering - not to pour budget into auto.
A high Top-of-Search modifier on auto will push spend toward the most expensive placement without knowing which auto segments convert there. That is a bid increase on unproven traffic.
Brand-name traffic (3%) is inside this campaign. It may be cannibalizing organic brand sales and inflating the ACOS.
2. Give a scaling plan that captures the upside without importing the waste.
Read the worked answer
Phase 1 - Harvest (week 1). Pull the search term report. Move every converting term (≥2 orders, CVR ≥10%) into a new manual exact-match campaign at a bid sized to a 25% ACOS target: break-even CPC ≈ $60 × 0.14 × 0.25 ≈ $2.10; bid below that.
Phase 2 - Prune auto (week 1). Add negative exact for the loose-match losers and for the brand name (run brand in its own campaign). Consider lowering the auto campaign bid on loose-match by reducing the base bid, since you cannot bid per match type in a single auto campaign - or split auto into two campaigns (close-match only vs. loose) if volume justifies it.
Phase 3 - Scale the proven (weeks 2-4). Scale budget on the manual exact campaign, not auto. Increase budget in 20-30% steps and watch ACOS and impression share. Add the Top-of-Search modifier only to the manual campaign, where you know the terms convert.
Phase 4 - Account for returns. Set the target ACOS to 25% gross / ~28% net of returns so the 12% return rate is baked in. Report on net sales, not gross.
Guardrails: pause scaling if net ACOS exceeds 30% or if loose-match share in auto rises above 15%. Keep auto funded at a modest discovery budget.