Amazon PPC Optimization: A Diagnostic Guide to Twelve Failures
Amazon PPC optimization starts with one question. Which number broke? Impressions, clicks, and orders form a funnel, and only one stage fails at a time. Zero impressions points to review status, bid floor, indexing, or budget. Clicks with no sales points to the listing, not the bid. A cost per click above your bid has a published cause, and so does a day that runs over budget. Find the broken stage, confirm it in a report, then change one thing.
Start here: which number broke
Every failure below sits at one of three stages. Your ad did not show. Your ad showed and nobody clicked. Someone clicked and nobody bought.
Check them in that order. A seller who raises bids on flat orders just buys more clicks on a page that fails to convert.
Pull two files first. The campaign report by day, and the search term report for the last 30 days. Almost every answer here sits in one of those two. For the mechanics, read how Amazon PPC works.
| Symptom | Likely cause | Report to check | Fix |
|---|---|---|---|
| New campaign, zero impressions | Still in moderation review | Campaign status | Wait for live status, then check the bid. |
| Zero impressions after two days | Bid under the auction floor | Suggested bid range | Raise to the low end of it |
| Impressions on some keywords only | Listing does not rank for the rest | Keyword plus ASIN in Amazon search | Rework title and backend terms |
| Impressions dropped overnight | Lost the featured offer | Detail page sales and traffic | Restore price or fix fulfillment |
| Impressions fine, click rate low | Main image, price, rating, or badge | Placement report | Fix offer and image before bids |
| Clicks, no sales, odd search terms | Match type pulls wrong traffic | Search term report | Add negatives, move winners to exact |
| Clicks, no sales, relevant terms | Offer loses the page comparison | Detail page sales and traffic | Price, reviews, content, delivery |
| Converting, then flat | New bidder, or a variation change | Campaign trend and search terms | Rebid top terms, recheck the parent |
| Cost per click above your bid | Dynamic bids plus placement adjustment | Placement report | Fixed bids, or cut the adjustment |
| Day spend above daily budget | Amazon averages budget over the month | Spend by day | Lower the daily figure |
| Sales look high for the ad spend | Attributed orders include other ASINs | Purchased product report | Judge on advertised ASIN sales |
| ACOS on target, bank balance down | Break even ACOS below reported ACOS | Your fee and cost sheet | Recalculate break even |
No impressions at all
Zero impressions has six causes, in order.
A new campaign sits in review before it serves. Check status first.
Next, the bid. A bid under the auction floor enters no auctions, so you record nothing. Move to the bottom of the suggested bid range and watch for 24 hours.
Third, indexing. No campaign can buy impressions for a keyword your listing does not rank for. Search that keyword plus your ASIN on Amazon. If your product does not appear, you have an Amazon SEO problem, not a bidding problem.
Fourth, the offer. Out of stock, or losing the featured offer, and Sponsored Products stops serving.
Last, the dull causes. A zero daily budget. An end date in the past. A category target with no matching products.
Impressions but almost no clicks
A low click rate is a listing problem. The auction already gave you the impression. The shopper read your thumbnail, price, rating, review count, and delivery promise, then picked someone else.
Raising the bid makes it worse. You buy more impressions of an ad shoppers skip, so cost per click climbs while orders stay flat.
Fix the visible parts in order: main image, price against nearby offers, review count, then rating. Our notes on listing images cover most of this.
Before you call your click rate low, compare it against the placement. Top of search behaves nothing like a product page slot.
| Placement | What this placement is | How to judge it | Note |
|---|---|---|---|
| Top of search, first page | Your highest intent slot | Against your own top of search history, never against another placement | Highest intent, highest cost per click |
| Rest of search | Volume traffic on broader matches | Against the same campaign’s top of search rate, then close the gap | Most volume, and most waste |
| Product pages | Shoppers comparing you against the ASIN they are already on | On price parity with that ASIN first, conversion second | Comparison traffic, judge on price parity |
Amazon publishes no placement benchmark. Set your baseline from your own placement report over the last 60 days, then judge every change against that.
Clicks but no sales
This splits into two cases, and the search term report tells you which.
Case one, the traffic is wrong. Read the actual shopping terms, not the keywords you typed. Broad match pulls related searches that never convert, such as a part for another brand. Add those as negative exact, then move the converting terms into their own exact match group.
Case two, the traffic is right and the page loses. Most end here.
Price. Open your own detail page and read the other offers on it. Shoppers compare in seconds.
Reviews. A product with nine reviews loses to one with nine hundred at the same price.
Content. Does the page answer the question the search term implies? Size, compatibility, material, quantity.
Stock and delivery date. A slow date beside a Prime competitor kills the order after the click.
Then variations. A split parent scatters reviews and pushes ad traffic to a child nobody wants.
It was converting and then it stopped
An abrupt stop almost always has an outside cause. Your campaign did not change, so the page or the auction did it.
First, the featured offer. Open detail page sales and traffic for the day the drop began. A fall in featured offer percentage explains most sudden stalls, because Sponsored Products stops serving without it.
Second, the auction. Compare average cost per click either side of the drop. A jump means a new bidder. Flat cost with fewer impressions means a lower rank.
Third, the listing. Look for suppression, a changed title, or a category move. Any of the three can drop a keyword’s index.
Fourth, inventory. Low stock slows the delivery promise and hurts conversion before it stops the ad.
Fifth, variations. A merge or a split moves reviews and changes which child collects the traffic. The first two explain most cases we see.
Cost per click higher than the bid I set
Two settings stack, and together they explain nearly every case.
Dynamic bids up and down let Amazon raise your bid. Amazon states it will “increase or decrease your bids by up to 100% for all placements based on performance” (dynamic bidding guide).
Placement bid adjustments sit on top. Amazon’s guidance says “You can add placement bid adjustments up to 900%” (Sponsored Products best practices).
Work the arithmetic on a one dollar bid.
- Base bid: $1.00
- Top of search adjustment of 100 percent: $2.00
- Dynamic uplift of up to 100 percent: $4.00 ceiling.
Push the placement adjustment to 300 percent and that ceiling reaches $8.00. Amazon publishes no worked example combining both settings, so treat this as ceiling arithmetic, not a formula.
You do not pay that ceiling on every click. Amazon puts it plainly: “Your final cost-per-click will be based on your adjusted bid plus additional factors, but will never exceed your maximum adjusted bid.” The ceiling caps you. It does not price you.
The fix sits in the placement report. If top of search costs three times the rest and converts no better, cut the adjustment first. Touch the base bid second.
Spend exceeded my daily budget
This one has a published answer that most sellers never read.
The budget best practices guide states: “Your daily budget is averaged over the course of a calendar month. That means you could spend less than or up to 25% more than your average daily budget on any given day.”
The same guide gives an example: “If your daily budget is $10, then your actual spend on a day could be $12.50 (25% higher).”
So a $10 budget with a $12.50 day matches the documentation, not a billing fault. Across a month that guidance describes total spend equalling $300.
Here the documentation stops. Amazon does not publish how it picks which days run hot, and it gives you no live ledger during the day. You see the overage afterwards.
The fix is arithmetic, not an appeal. Set the daily figure at a level you can accept 25 percent above. For a hard stop in promotion week, use a budget rule or a scheduled pause.
My ROAS includes sales I did not make
Two habits make reported return on ad spend look better than your bank statement.
The first has documentation behind it. Amazon defines the search term report as the shopping terms “that resulted in at least one click on your ad.” The click is the trigger. What the shopper buys next decides which order lands in the report, and it need not be the product you advertised. Amazon publishes a purchased product report that splits attributed purchases by the ASIN the shopper bought.
So an ad for one ASIN can collect credit for a different ASIN in your catalog. Campaign ROAS looks strong. The advertised product can still lose money.
The second case needs care, and here we reason rather than cite. On a listing shared by several sellers, the featured offer can change between the click and the checkout. Amazon does not publish, in one plain page, how attribution behaves when the shopper buys that same ASIN from another seller. We do not claim to know, and anyone who answers confidently should show you the source.
Do this with both. Pull the purchased product report monthly. Compare advertised ASIN sales against total attributed sales. A wide gap means your campaign number measures your catalog, not your ad.
ACOS is fine but profit is not
Reported ACOS is a ratio. On its own it says nothing about profit. The number that does is break even ACOS, which is contribution margin divided by selling price.
Two examples show the method. Run yours from your own fee sheet.
Product A sells for $30. Referral fee, fulfillment, and cost of goods total $24. Contribution is $6, so break even ACOS is 20 percent. At a reported 25 percent ACOS, every advertised order loses money.
Product B sells for $60. The same three costs total $30. Contribution is $30, so break even ACOS is 50 percent. At 45 percent ACOS, every advertised order pays.
A 25 percent ACOS is the worse number here. Any blanket rule such as “cut anything above 30 percent” quietly destroys profit on Product B.
Then widen to total advertising cost of sales, ad spend divided by all revenue. ACOS grades the ad. TACOS grades the business. Climbing TACOS with steady ACOS means ads now carry sales organic used to win free.
Campaign structure that prevents most of this
One structure, four layers. Each answers a different question, and mixing them creates data you cannot read.
Single keyword ad groups earn their place only where volume supports them. Below that, they multiply admin without improving a decision.
| Layer | Purpose | Match types | Bid approach | Share of budget |
|---|---|---|---|---|
| Discovery | Find new search terms | Broad, phrase, auto | Low fixed bids, wide net | 20 percent |
| Performance | Scale proven terms | Exact | Bid to break even ACOS | 55 percent |
| Competitor and product | Win the comparison slot | ASIN and category targets | Fixed and capped | 15 percent |
| Defensive | Hold your own brand terms | Exact, brand targeting | Low, enough for top of search | 10 percent |
That split is a starting point, not a rule. Discovery gets a fixed share so new terms keep arriving without draining the campaigns that pay the bills. Performance takes the majority because it holds terms you already proved. Defensive stays small, since brand traffic converts well at a low bid.
The weekly routine
Set a cadence and stop editing every morning. Bid changes need days of delivery before they mean anything.
| Frequency | What to check | Where | Action threshold |
|---|---|---|---|
| Daily | Spend pace, stock, capped campaigns | Campaign manager | Budget capped early, or an ASIN out of stock. |
| Weekly | Search terms, negatives, placement cost | Search term and placement reports | A term past break even spend with no order. |
| Monthly | ACOS against break even, TACOS, budget split | Reports and your cost sheet | Any layer ten points off share |
Make the changes in bulk. The console drags past a handful of edits. Download the bulk file, filter, edit the bid and status columns, then upload. Bulk files also record what changed and when.
Dayparting pays when hourly data shows a pattern repeating across several weeks and your spend makes a small saving worth the work. Below that, it costs impressions in hours you guessed wrong about.
Mistakes that look like Amazon PPC optimization
Five habits that feel productive and cost money.
Cutting bids on every high ACOS keyword. A term with four clicks and no order tells you nothing. Check click volume before you touch the bid.
Adding negatives too early. Each negative shuts a door for good, and you never see the orders you stopped receiving.
Pausing a campaign in its first days. New campaigns need a stretch of steady delivery before their numbers settle.
Judging on too few clicks. Pick a click threshold tied to your conversion rate and hold to it.
Chasing the lowest possible ACOS. Cut deep enough and the ratio looks excellent while profit falls. Profit buys your next purchase order. A ratio does not.
Frequently asked questions
Why am I getting clicks but no sales on Amazon PPC?
Two causes. Either the search terms are wrong, or the listing loses the comparison on the page. Open the search term report first. If the terms match your product, the problem sits in price, reviews, images, content, or delivery. Fix the page before you raise bids.
Why does my campaign have zero impressions?
Check six things in order. Campaign still in review. A bid under the auction floor. A listing that does not rank for the keyword. Out of stock or a lost featured offer. A zero budget. An end date in the past. Test indexing with the keyword plus your ASIN.
Why did my Amazon campaign stop converting?
Something outside the campaign changed. Check featured offer percentage on the day of the drop. Then average cost per click, for signs of a new bidder. Then the listing for suppression or a title edit, then stock, then variations. The first two explain most stalls.
Why is my cost per click higher than my bid?
Dynamic bids up and down can raise a bid by up to 100 percent. Placement bid adjustments go up to 900 percent on top. A one dollar bid with a 100 percent top of search adjustment reaches a four dollar ceiling. Amazon says your final cost per click never exceeds your maximum adjusted bid.
How did my campaign spend more than the daily budget?
Amazon averages your daily budget across a calendar month and allows up to 25 percent more on any single day. A ten dollar budget can spend twelve dollars fifty in a day. That matches Amazon’s published guidance rather than a billing error. Set the daily figure accordingly.
How many clicks before I judge a keyword?
Tie the threshold to your conversion rate, not a fixed number. If your page converts one order in ten clicks, ten clicks with no order means almost nothing. Wait until you have enough clicks that two or three orders would look normal. Four clicks is guesswork.
When should I add a negative keyword?
Add it when a term has real click volume, no orders, and no plausible route to one. Clearly irrelevant terms can go straight away. Relevant terms with thin numbers deserve patience. A negative closes that door for good, and you never see the orders you lost.
How do I undo a negative keyword?
Remove the term from the negative list in the ad group or campaign where you added it, or archive that entry. Check both levels, since a campaign level negative blocks every ad group inside it. Delivery does not resume the same hour, so wait a few days.
Should I advertise every variation or only the best sellers?
Advertise the children that convert and let the rest run on organic traffic. Spend spread across every size and color buys clicks on variations shoppers skip. Open detail page sales and traffic by child ASIN, pick the two or three carrying sales, and build exact groups there.
Is dayparting worth setting up?
Only with an hourly pattern that repeats across several weeks, and spend large enough that a small saving covers the work. Smaller accounts usually lose more from missed impressions than they save. Check the hourly data, confirm the pattern holds, then schedule.
How often should I change bids?
Weekly suits most accounts. Bid changes need days of delivery before the data means anything, so daily edits leave you reading noise. Move in steps of ten to twenty percent, change one thing at a time, and log every edit. Next week’s number then has a cause.
Does pausing a campaign reset its performance?
Amazon publishes no rule saying a paused campaign loses its history or restarts a learning period. What we see in practice is a slower ramp once delivery resumes. Treat a pause as a real cost. To slow a campaign instead, cut the budget or the bid.
Start with the symptom, not the settings
Pick the row that matches what you see, run the check beside it, then apply one fix. Confirm the answer in a report before you change a setting.
To hand the diagnosis to a team that does this daily, our Amazon PPC management service starts with a free audit, one named account manager, and a written report every month. Your files and logins stay yours.
