A Budget Increase That Bought the Wrong Placements
Your target ACOS is 30%. Campaign 'Trail Bottles' uses dynamic bids - down only, an $80 daily budget, and a +50% top-of-search placement modifier. It exhausts its budget by noon, holds 35% impression share on its main keywords, and reports a 22% ACOS on top-of-search placements against 48% on product pages and rest of search. A second campaign, 'Trail Lids,' has a $40 budget it never fully spends, 8% ACOS, and 12% impression share. You raise Trail Bottles to $200 and switch it to dynamic bids - up and down; two weeks later ACOS is 38%, spend still ends by 1 p.m., and top-of-search ACOS is unchanged at 22%.
1. Was the budget increase plus bidding-strategy switch the right lever? Decompose what actually happened.
Read the worked answer
No - you treated a mix problem as a size problem.
The campaign was already reporting two very different businesses: top of search at 22% (better than your 30% target) and product pages / rest of search at 48% (far worse). When a campaign is budget-capped, the marginal dollar goes to whichever eligible placement is available late in the day - here, the expensive placements. Raising the budget from $80 to $200 therefore bought more of the 48% inventory, and blended ACOS moved from roughly 30% territory to 38%.
The bidding switch compounded it. Dynamic bids - up and down raises bids where Amazon predicts a higher chance of conversion, and with a hard budget the campaign simply burns faster at a higher CPC. Top-of-search ACOS staying at 22% confirms the extra spend did not go there.
The lever that fits the data: ACOS = CPC / (conversion rate x average order value). You need a mix shift, not more money - raise the top-of-search modifier (modifiers can go as high as 900%) so more of the same budget competes for the 22% placement, and lower the base bid so the 48% placements clear less often. Net spend can stay flat while the blend improves.
2. What is the right move for 'Trail Lids,' and what specifically would you check before making it?
Read the worked answer
Trail Lids is under-invested, not underperforming. An 8% ACOS against a 30% target means roughly three times more cost per order is affordable. Twelve percent impression share with an unspent budget tells you the constraint is not money - if it were, the budget would be gone.
Check in this order:
- Eligibility - Buy Box ownership, stock status, and whether the product is suppressed or missing required attributes. Low impression share with unspent budget and a great ACOS is frequently a listing problem, not a bidding problem.
- Bid floor - if eligible, raise base bids in steps (15-20% at a time, two weeks apart) and watch impression share respond.
- Placement - add a top-of-search modifier once bids are competitive, since top of search typically converts better than product pages and you have headroom.
- Relevance - if impression share stays flat as bids rise, the keyword-to-product relevance or the listing's quality signals are capping you, and more money will not help.
Do not reallocate the Trail Bottles increase into Trail Lids as an offset; solve the Bottles mix first so you are comparing two clean campaigns.
3. Two weeks after the change ACOS is 38%. Walk through how you would diagnose it before making another change, including any measurement traps.
Read the worked answer
Split ACOS into its parts rather than reacting to the headline.
- Placement report: did the extra $120 land on product pages and rest of search? If yes, the diagnosis is confirmed and the fix is modifier up, base bid down.
- CPC vs conversion rate: if CPC rose while conversion rate held, the bidding switch is the cause; revert to dynamic bids - down only. If CPC held and conversion rate fell, you are reaching less-qualified traffic and should tighten targeting rather than bids.
- Impression share and lost-impression data: distinguish lost to budget from lost to rank. If you are still losing impressions to budget at 1 p.m., the campaign is still size-constrained - but that does not mean the added size was well spent.
- Measurement traps: the recent week is understated because conversions credit to the click date within the attribution window, so a two-week read partly measures lag. Two weeks is also thin for a conclusion; compare on equal windows and prefer a four-week read against a matched prior period.
- Seasonality and price: a price change, coupon, or competitor move in the same window can produce the same 16-point ACOS swing. Check the detail page and the buy-box price for changes before attributing the result to the bid change.