Budgets and placements — the two levers most people ignore

Bidding gets the attention because it touches every keyword. But two other levers move accounts just as hard, and both are routinely left untouched for years.

Part one: the campaign switching itself off

A campaign that reaches its daily budget stops serving for the rest of the day. Not throttled — stopped.

Amazon does not usually spread the budget evenly. A campaign with a $50 cap that spends $50 by 3pm shows nothing from 3pm onward. It misses the evening entirely, which for most categories is when conversion is strongest. You bought the worst hours of the day and skipped the best ones.

Now consider a campaign doing that while beating its ACOS target.

It is profitable. Every dollar it spends returns more than you asked for. And it switches itself off every single day, at the point where it would start doing its best work.

This is the clearest signal in Amazon advertising and the most commonly missed one, because nothing in the console shouts about it. A campaign under target looks healthy, and "budget utilisation 100%" reads like efficiency rather than a constraint.

Raising bids here makes it worse. The campaign is not short of auctions to win, it is short of money. Higher bids spend the same budget faster and cap out earlier in the day. The correct lever is the budget.

How to raise a budget without losing control

Not every capped campaign deserves more money. Four conditions worth requiring:

It is genuinely efficient, not merely capped. Capped and over target means the bids are wrong; more budget just loses money faster.

It caps consistently. One busy day is not a pattern. Several days out of seven is.

There is stock to sell. Buying more traffic for something about to run out wastes the spend and damages the listing's momentum when it goes unavailable.

There is an account-level ceiling. A per-campaign rule with no global cap can walk your total spend up indefinitely, one defensible increase at a time.

Raise in steps rather than doubling. A large jump changes how Amazon paces the campaign and you lose the ability to tell whether the extra money bought anything.

Cutting a budget is a stopgap, not a fix

The reverse case: a campaign spending far beyond its target with real money at stake.

Cutting the budget works immediately, which is exactly why it is tempting. Bids take time to re-price traffic; a budget cut takes effect today.

But it treats the symptom. The campaign is not overspending because its budget is too high — it is overspending because its bids are wrong for what it earns. Cut the budget and you have capped the damage while leaving the cause running, now inside a smaller box.

Use it to stop bleeding. Then fix the bids. And do not let it auto-restore — putting the budget back is a decision you make once the underlying problem is actually solved.

Part two: why lowering bids raised your ACOS

Your effective cost per click is not your bid. It is:

Effective CPC = base bid × (1 + placement modifier)

Amazon lets you add a percentage for top of search and for product pages. At a $1.00 base bid, a 15% top-of-search modifier means $1.15 there and $1.00 everywhere else.

This is why the most common complaint in Amazon advertising has a mechanical explanation.

"I lowered my bids and my ACOS went up."

ACOS is CPC ÷ revenue per click. If CPC fell and ACOS rose, revenue per click must have fallen further. Conversion got worse. Two usual causes:

Your placement mix changed. A lower bid may change where ads appear. Check the realized placement report: top of search is not always the best-converting placement, and product pages are not always the worst. Compare the same dates and scope before attributing a change to placement.

You lost your best search terms. On broad, phrase and auto targeting, different queries clear at different prices — and the expensive ones are frequently the good ones. Lower the bid and you keep only the cheap, poorly-converting tail.

What to aim for

Roughly equal ACOS across placements, with spend concentrated where conversion is strongest.

If top of search runs 15% ACOS and product pages run 60%, investigate the sample size, stock, current modifiers and goals. A modifier changes the auction bid, not realized CPC. There may be room to shift traffic, but neither unchanged total spend nor improved ACOS is guaranteed. Sequence eligible changes and observe the result.

Two cautions

Only tune placements on realized placement data. You need the campaign × placement × day report. Inferring placement performance from search-term breadcrumbs produces a modifier that is a guess wearing a number.

Move one lever at a time. Bid and modifier together mean you cannot attribute the result to either. Next week you will have spent a week of evidence and learned nothing. Pick one, move it, wait for attribution to settle, then judge.

Both levers have a floor of patience

Budget and placement changes need the same settling time as bids. Amazon attributes sales to the click that caused them, and shoppers do not buy instantly — the first few days after any change look worse than reality because the spend has landed and some of the sales have not.

Judging inside that window is how good changes get reverted.

What AutoPPC does with this

Budget raises require target efficiency, a consistent capping pattern, verified stock cover and your account daily cap. Brakes are explicitly labelled as stopgaps whose reason states that bids are the underlying fix, and they never auto-reverse.

Placement changes are proposed only from realized campaign × placement × day evidence, with a grace band and a share cap so raises stop when a placement's spend share reaches your limit.

Bid and placement modifier never move on the same campaign in the same run — an Invariant, not a setting, because a change you cannot attribute is a change you cannot learn from.

Source: Amazon Sponsored Products bidding and placement guidance.