Amazon Sponsored Ads common mistakes
Amazon Sponsored Ads · Understand the traps and how to avoid them.
Launching broad- and phrase-match keywords without adding negative exacts for the terms already covered by an exact-match keyword elsewhere in the account, then assuming Amazon routes each search term to the single 'best' keyword.
Candidates treat negative targeting as cleanup for obviously irrelevant queries only, and assume the system automatically resolves overlap between match types so duplicate coverage is harmless.
What to do instead
Build a canonical exact-match keyword set first, then add those same terms as negative exacts inside every broader-match campaign. Review the search-term report on a fixed weekly cadence and promote winners to exact while negating them in the broad campaign.
Cutting a keyword's bid as soon as its ACOS exceeds target, without first checking whether the spend sits in a placement (top of search, product pages, rest of search) with a materially worse conversion rate.
Candidates read a single blended ACOS number and jump to the bid field as the only lever, forgetting that placement modifiers and match-type mix change the economics of the same keyword.
What to do instead
Segment the keyword by placement before touching the bid. If the loss is concentrated in one placement, lower that modifier rather than the base bid; only reduce the base bid when performance is weak across placements, and re-check after the learning period.
Comparing an ACOS pulled from a report defaulted to a 1-day attribution window against an ACOS taken from a 14-day figure in a different report, and treating the gap as a real performance change.
Candidates do not check which attribution window and date range a report was generated under, and assume all 'attributed sales' columns are directly comparable across report types.
What to do instead
Fix the attribution window and date range explicitly before any comparison, document the setting next to each metric, and only compare like-for-like. When evaluating a change, compare the same window before and after rather than mixing defaults.
Concluding a campaign is 'fully funded' and has no delivery constraint because spend equals the daily budget, then searching for a bidding explanation for flat impressions.
Candidates treat the daily budget as a hard ceiling and as proof of sufficient funding, missing that spend can run above the daily figure on individual days while averaging out over the month, and that budget caps still suppress impression share.
What to do instead
Check out-of-budget and lost-impression-share indicators alongside spend, and look at budget utilisation across the month rather than a single day. If the campaign is budget-capped, raise or reallocate budget before adjusting bids.
Planning a Sponsored Brands Video launch for a brand that is not enrolled in Brand Registry and has no registered or pending trademark, and reusing a long landscape brand film as the creative.
Candidates generalise from Sponsored Products eligibility, which is far broader, and assume Sponsored Brands and Sponsored Brands Video accept the same assets and requirements as each other.
What to do instead
Verify Brand Registry enrolment and an active professional selling account before designing the campaign, then check the current Sponsored Brands Video creative spec for duration, aspect ratio, file size, and required elements. Recommend an alternative format if eligibility is not met.
Building every Sponsored Display campaign as views-remarketing retargeting of people who already viewed the product detail page, and never testing Amazon-audience or product-targeting tactics.
Remarketing is the most familiar Sponsored Display use case, so candidates default to it and treat Sponsored Display as a single-tactic channel rather than a set of distinct audience and targeting options.
What to do instead
Match the tactic to the funnel stage: product and category targeting for prospecting, in-market and interest audiences for consideration, and remarketing for conversion. Launch comparisons for each and judge them on their own cost-per-acquisition targets.
Placing dozens of unrelated ASINs or keywords into one ad group under a single bid, so that low-price and high-price products, or head and tail terms, all share one economics setting.
Candidates think of ad groups as folders for tidy organisation and want fewer campaigns and ad groups to manage, underestimating how a shared bid destroys the ability to control spend at the level that actually converts.
What to do instead
Structure ad groups around a coherent theme with comparable price points and conversion behaviour so one bid is defensible for the whole group. Split out outliers into their own ad group, and use the search-term and advertised-product reports to confirm the grouping still holds.
Submitting an ad for a product whose detail page has a weak star rating, an incomplete or non-compliant main image, sparse bullet points, or a recurring out-of-stock history, and then diagnosing the resulting low conversion rate as a targeting problem.
Candidates focus on the ads console and treat the detail page as someone else's responsibility, forgetting that click-to-conversion depends on retail readiness the ad cannot fix.
What to do instead
Run a retail readiness check before launch: in-stock and Buy Box eligibility, review rating and count, compliant main image, complete title and bullets, A+ or brand content where available. Fix the listing or hold the spend, then re-baseline performance after the changes.
Reacting to a week-over-week sales drop by immediately lowering bids, without first checking Buy Box loss, stock status, budget-capped days, or impression share lost on the campaign.
Bid is the most visible lever in the interface, so candidates reach for it first and skip the sequential diagnostic that separates supply and eligibility problems from genuine bidding problems.
What to do instead
Diagnose in order: is the product in stock and winning the Buy Box, is the campaign budget-capped, has impression volume changed, has click-through or conversion rate changed. Only adjust bids once the upstream causes have been ruled out, and change one variable at a time.
Setting product targeting on whole categories or on competitor ASINs and leaving it at that, with no refinement by price, rating, or Prime eligibility and no negative product targets added from the report.
Candidates select a target and consider the campaign finished, treating product targeting as a single decision rather than an ongoing process of narrowing reach as the report reveals which targets spend without converting.
What to do instead
Refine category targets with the available price, rating, and delivery filters at launch, review the targeting report for spend without conversion, and add negative product targets for ASINs that never convert. Re-evaluate the refined set on a scheduled cadence.