How to structure Sponsored Products campaigns on Amazon

The messy Sponsored Products accounts we come across nearly always share the same problem: campaigns created one at a time, with no common logic, where the same keyword competes with itself in three different places. The result is that nobody knows what’s working any more. Here we explain how to set up a goal-based structure, how to use match types and negatives, and how often to optimise.

Why structure matters more than bids

A good structure lets you answer simple questions: how much am I spending to be found by people searching for my brand? How much am I investing to win generic searches? Which new keywords are emerging? If your campaigns are mixed up, you don’t have those answers, and every bid change becomes a shot in the dark.

The rule we follow is simple: every campaign has one goal, and every important search term lives in one place only.

A fixed naming convention also helps a lot, for example product, product family, targeting type and match type. It sounds like a detail, but once you have a hundred campaigns it’s the only way to filter reports and see at a glance where the budget is going.

Four campaign families, by goal

1. Research

These are for discovering search terms and ASINs that convert. This is where automatic campaigns and manual broad match campaigns sit. Modest budget, moderate bids, goal: collect data.

2. Scaling

This is where keywords that have proven they sell end up, in exact match. These are the campaigns where you really push, with higher bids and bigger budgets. Keep generic category terms separate, as they have different volumes and costs.

3. Brand defence

Campaigns on searches containing your brand name and on your own product pages. They’re cheap per click and convert well, but read them carefully: some of those sales would have happened anyway. Their job is to stop competitors taking space on your pages and your searches.

4. Competitor

Product targeting on competitors’ ASINs and keywords with their brand names. They work when you have a clear advantage to show (price, reviews, pack size). Measure them separately, because they usually have a higher cost per sale.

Automatic and manual: making them work together

Automatic campaigns let Amazon choose where to show the ad, based on the listing content. You can manage them through four targeting groups: close match, loose match, substitutes and complements. It’s worth splitting them out, or at least setting different bids, because they perform very differently.

Manual campaigns give you control: you choose keywords or products and set the bids. The typical flow is:

1. The automatic or broad campaign finds a term that sells.
2. The term moves into a manual exact match campaign, in the scaling family.
3. The same term is added as a negative exact in the research campaigns, so it doesn’t compete with itself.

Repeated consistently, this step is what turns a confused account into one that improves over time.

Match types and product targeting

In product targeting, Amazon also offers an “expanded” option that extends the ad to ASINs similar to the one you chose. It can help with discovery, but in competitor campaigns check that it’s set the way you want.

Negative keywords: half the job

Every week the search term report shows searches that spend without selling or have nothing to do with the product. They need excluding. Negatives can be phrase or exact, and you can also exclude specific ASINs.

• Relevance negatives: terms describing a different product (for example a format or use you don’t offer).
• Performance negatives: relevant terms that, with enough data, don’t convert at an acceptable cost.
• Structural negatives: terms already moved to exact, excluded from research campaigns.

Be careful not to overdo it: excluding a term with only a few clicks too early starves the campaigns that need that data.

Budgets, bids and placements

Daily budget shouldn’t run out halfway through the day in your scaling campaigns. If it does, you lose evening sales, which are often the best. Better to lower bids on what performs worse than leave good campaigns without budget.

For bids, Amazon offers several strategies: dynamic bids (down only), dynamic bids (up and down) or fixed bids. In research campaigns, fixed or down-only bids keep costs under control. In scaling campaigns with solid data you can give the algorithm more freedom.

You can also raise bids by placement: top of search, rest of search, product pages. Look at the placement report before you do: the top of the first page converts better, but it also costs more.

For a wider view of how we set up the whole advertising plan, see our Amazon Advertising services page.

Optimisation rhythm

• Daily: only check for exhausted budgets and obvious anomalies.
• Weekly: search terms, negatives, moving winning terms to exact, bid adjustments.
• Monthly: review the structure, the share of budget across the four families, and compare with total sales and TACoS.

Ad sales are attributed with a delay of a few days, so judging a change after 24 hours nearly always leads to the wrong decision. To read spend, sales and margin together we use Bettershop Analytics, so campaign decisions take organic sales into account too.

Common mistakes

• A single automatic campaign for the whole catalogue.
• Products with very different prices and margins in the same ad group.
• The same keyword in several campaigns with no cross-negatives.
• Pushing ads on products that are out of stock or have few reviews.
• Judging every campaign against the same target ACoS, brand and competitor included.

What to do now

Export your list of active campaigns and assign each one to one of the four families. Those that don’t fit any family, or fit more than one, are the first to rebuild. Then download the search term report for the last 30 days: move terms that sell to exact and add the ones that spend without results as negatives.

Frequently asked questions