Sponsored Display audiences let you show ads to shoppers based on their behaviour rather than their search terms, and the single biggest lever you control is matching the lookback window to how long people actually take to buy your product. Use audiences to prospect new shoppers or win back browsers and buyers; use contextual or product targeting when you want to sit next to specific listings. Measure new-to-brand share first, because it tells you whether the audience is actually expanding your customer base.
TL;DR:
- Audience lookback windows should align with your product’s purchase cycle, such as 7 days for fast-moving goods and 14 to 30 days for considered products.
- Using Amazon Audiences and dynamic segments is best for brand awareness, while views and purchases remarketing target consumers closer to making a purchase.
- Combining off-Amazon placements through remarketing with on-Amazon inventory improves reach, especially for audiences who have not yet bought from you.
- Automated creative functions well for testing broad audiences but custom images can outperform generic templates in remarketing campaigns.
- Regularly monitoring metrics like new-to-brand share and cost per acquisition helps optimize audience targeting and avoid wasteful spend.
Table of Contents
- The main types of Sponsored Display audiences
- Where your audience ads actually appear
- Matching lookback windows to the buying cycle
- Optimised targeting versus manual control
- Billing models and choosing the right optimisation objective
- What to measure and how to read the numbers
- An actionable setup checklist
- How analytics and automation speed up audience testing
- How to create and manage custom audiences for Sponsored Display campaigns
- Best practices for audience segmentation and layering
- Case studies or examples demonstrating effective audience targeting
- Limitations and policies related to Sponsored Display audiences
- Integrating Sponsored Display audiences with Sponsored Brands and Sponsored Products
- Where audience targeting fits in a full-funnel strategy
- Getting more from your Sponsored Display audience tests with Osellpa
- FAQ
- Sources
The main types of Sponsored Display audiences
Each audience type pulls from a different signal, so picking the right one depends on where the shopper sits in their decision.
- Amazon Audiences: pre-built lifestyle, interest and in-market segments that Amazon assembles from shopping and browsing signals, useful for reaching cold shoppers who have shown relevant intent but have not yet seen your product.
- Views remarketing: targets shoppers who viewed your product detail page (or similar products) without purchasing, which makes it one of the more direct ways to recover an almost-sale.
- Purchases remarketing: targets shoppers who already bought your product or a related one, and can be refined by price band, brand or star rating to sharpen who sees the ad.
- Dynamic segments: machine-learning groups such as “audiences likely to be interested in your ad,” which Amazon recommends layering alongside Amazon Audiences, views and purchases to add reach without manual list-building, according to Amazon Ads guidance on dynamic segments.
- AMC audiences: rule-based or lookalike segments built in Amazon Marketing Cloud that extend Sponsored Display beyond the standard pre-built options, letting advertisers combine cross-channel signals that a normal segment cannot capture, as described in Amazon’s guidance on AMC audiences for sponsored ads.
Amazon Audiences and dynamic segments generally serve awareness and consideration, where the goal is to get in front of relevant shoppers who do not know your brand yet. Views and purchases remarketing serve the bottom of the funnel, where the shopper already knows the product and needs a nudge or a reason to come back.
Where your audience ads actually appear
Placement depends partly on which audience and targeting type you choose, and it changes what creative you should prepare.
On Amazon, Sponsored Display ads can show on product detail pages, search results and customer review pages. Off Amazon, they can appear on third-party apps and websites through Amazon’s ad network, which is where remarketing audiences tend to deliver the most incremental reach.
- Views and purchases remarketing lean heavily on off-Amazon placements to recapture attention away from the marketplace.
- Amazon Audiences and dynamic segments mix both on-Amazon and off-Amazon inventory depending on where the matched shopper happens to be browsing.
- Automated creative pulls your existing product images and copy to build the ad automatically, which works well when you are testing several audiences at once.
- Custom image creative gives you more control over messaging and branding, and tends to suit remarketing campaigns where a tailored offer or reminder outperforms a generic template.
Dynamic creative optimisation, available through self-service Sponsored Display, automatically serves whichever creative version performs best when a campaign promotes multiple products, according to Amazon’s display ads guide.
Matching lookback windows to the buying cycle
The lookback window decides how far back Amazon looks to decide whether someone qualifies for your remarketing audience, and getting it wrong wastes spend on shoppers who have already moved on or forgotten why they clicked.
Amazon’s own guidance recommends matching the window to how long a typical customer takes to decide, rather than defaulting to the same setting across every campaign, according to Amazon’s guide to views remarketing and lookback windows.
- Fast-moving goods (snacks, household basics, low-cost accessories) tend to suit a 7 day window, since the decision happens quickly and a longer window just adds stale clicks.
- Considered purchases (electronics, furniture, higher-ticket tools) often need 14 to 30 days or longer, because shoppers research before buying and a short window misses most of them.
- Mid-range products sit between the two: start at 14 days and shorten or lengthen based on what your conversion data shows after a few weeks.
- Test in pairs: run the same creative and audience with two different windows for two to three weeks, then compare cost per conversion and new-to-brand share before settling on one.
Pro Tip: Resist defaulting to the longest available window just to maximise reach. A wider window pulls in shoppers who were never close to buying, which inflates spend without improving conversion.
Optimised targeting versus manual control
Optimised targeting is Amazon’s dynamic version of Sponsored Display audiences: instead of you picking a fixed segment, Amazon continually shifts delivery toward whichever audience signals are converting. It is designed for Sponsored Display Conversion campaigns and can genuinely save time on manual segment management.
The trade-off is that optimised targeting can quietly stop serving a manual segment it judges inefficient, even if that segment is still valuable to you for brand reasons, according to Amazon’s documentation on optimised targeting for Conversion campaigns. Manual remarketing or contextual targeting tends to outperform automation when you need predictable delivery to a specific group, such as past purchasers of a particular variant, or when you are protecting a niche audience that is small but high-value.
- Weekly audit: export segment performance and compare cost per acquisition, click-through rate and new-to-brand share across manual and optimised segments.
- Two-week rule: if a manual segment shows poor conversion after two weeks, pause or lower its bid rather than letting it drain budget indefinitely.
- Hybrid approach: seed a campaign with manual segments you know perform, then layer in optimised targeting once you have baseline data to compare against.
Pro Tip: Keep at least one manual segment running permanently as a benchmark, even after optimised targeting takes over most of the budget, so you always have a control to measure against.
Billing models and choosing the right optimisation objective
Sponsored Display supports two billing models, and the right choice depends on what the campaign is actually meant to achieve.
- CPC (cost-per-click) suits performance-driven campaigns where you want clicks that lead to conversions, and it is the natural fit for views and purchases remarketing.
- vCPM (viewable cost-per-thousand-impressions) suits awareness campaigns where the goal is viewable impressions rather than clicks, which fits top-of-funnel Amazon Audiences or dynamic segment campaigns.
- Optimisation objectives should match the goal: choose conversions when the aim is sales, detail page views when the aim is consideration, or viewable impressions when the aim is pure reach.
- Self-service Sponsored Display campaigns carry no minimum spend requirement, which makes it practical to test several audience and billing combinations in parallel, according to Amazon’s optimisation guide for Sponsored Display.
As a rough starting point, set a lower CPC bid on broad Amazon Audiences and a higher one on purchases remarketing, since the latter audience is smaller and further along in the decision, so each click is worth more.
What to measure and how to read the numbers
A handful of metrics tell you almost everything you need to know about whether an audience campaign is working.
- New-to-brand (NTB): the share of sales coming from customers who have not bought your brand in the past 12 months, which is the clearest signal of genuine prospecting success.
- Conversions and detail page views: show whether the audience is moving people toward a purchase decision, not just clicking.
- ROAS and CPA: tell you whether the spend is efficient enough to justify scaling the audience.
- CTR: a quick diagnostic for creative relevance, though it should never be read alone.
Advertisers using a mix of Amazon Audiences, views and purchases remarketing have observed new-to-brand shares as high as 82% on some campaigns, according to Amazon’s lookback window guidance, which shows how much of Sponsored Display’s value sits in reaching shoppers outside your existing customer base.
Check performance weekly through Amazon’s standard reporting, and use AMC when you need deeper attribution across touchpoints. Treat high reach with flat conversion, or a high CPA on a remarketing segment that should be cheap to convert, as a red flag worth investigating immediately.
An actionable setup checklist
Running a new audience campaign is easier when you follow the same sequence every time.
- Set the objective first: awareness, consideration or remarketing decides everything downstream.
- Choose the audience type that matches that objective, from Amazon Audiences through to AMC lookalikes.
- Set the lookback window based on your product’s buying cycle, not a default setting.
- Prepare creative: automated for broad tests, custom images for targeted remarketing.
- Pick the billing model and optimisation objective: CPC and conversions for performance, vCPM and viewable impressions for awareness.
- Define the measurement window, typically two to three weeks before making changes.
- Run a control versus variant test: same budget, same creative, different audience or window, for at least two weeks.
- Apply refinements: price band, star rating or brand filters on purchases remarketing once you have baseline data.
How analytics and automation speed up audience testing
Testing audiences faster, without burning budget on the way, usually comes down to how quickly you can see which segment and bid combination is working.
- Profit tracking tied directly to Amazon sales data shows the real margin behind each audience test, not just the headline ROAS.
- Automated PPC bid optimisation reports flag underperforming segments and bid levels without a seller manually trawling spreadsheets.
- Automated alerts can surface a sudden spend spike or conversion drop on a specific audience before it eats the week’s budget.
A practical workflow: pull your Amazon Sponsored Display reports, run them through an automated bid optimisation report, then adjust lookback windows or bids based on what the report flags rather than waiting for a full reporting cycle to end.
How to create and manage custom audiences for Sponsored Display campaigns
Custom audiences in Sponsored Display are built inside the Amazon Ads console, under the Sponsored Display campaign creation flow, rather than uploaded from an external list.
Start by selecting the “Audiences” targeting type when creating a campaign, then choose from views remarketing, purchases remarketing or an Amazon Audience segment. For purchases remarketing, apply refinements such as price range, brand or star rating to narrow the pool to shoppers most likely to convert again. For views remarketing, decide whether to include views of your own products only or views of similar products, since the latter widens reach but dilutes intent.
Once live, manage the audience by reviewing performance at the ad group level weekly. If a segment is underperforming, you can lower its bid, pause it or swap it for a dynamic segment without rebuilding the whole campaign. Multiple audiences can run within the same ad group, which lets you compare several at once under identical creative and budget conditions.
Keep naming conventions consistent across ad groups, such as labelling by audience type and lookback window, so performance comparisons stay clear as the number of active audiences grows. Revisit refinements every few weeks, since a price band or star rating filter that worked at launch may need loosening or tightening as your catalogue or reviews change.
Best practices for audience segmentation and layering
Layering audiences well means giving each segment a distinct job rather than overlapping them all on the same goal.
Start broad with Amazon Audiences or dynamic segments to build awareness among relevant but unfamiliar shoppers, then narrow to views remarketing for people who showed direct interest, and finish with purchases remarketing for past buyers who are likely to repurchase or trade up. Running all three at once, with separate budgets, lets you see which stage of the funnel is actually driving sales rather than guessing from blended numbers.
Avoid stacking too many refinements on a single segment, since a purchases remarketing audience filtered by price band, brand and star rating all at once can shrink to a handful of shoppers, too small to deliver meaningful data. Behavioural segmentation concepts, such as grouping by browsing habits rather than demographics alone, can sharpen how you think about which Amazon Audience segments are worth testing first, as explained in this guide to behavioural targeting.
A useful rule: keep no more than two or three refinements active on any one segment, and test them one at a time so you know which change actually moved performance. For broader thinking on how to structure segments across a funnel, a wider audience segmentation guide covering behavioural and predictive approaches can help frame the decision before you build campaigns in the Ads console.
Case studies or examples demonstrating effective audience targeting
A straightforward example: a seller of a mid-priced kitchen gadget runs views remarketing with a 14 day window alongside purchases remarketing with a 30 day window for accessory buyers. The views segment recovers shoppers who clicked but did not buy within two weeks, while the purchases segment targets owners of a related product likely to need a refill or upgrade.
Another common pattern involves a seller testing Amazon Audiences for a new product launch, running vCPM bidding to build awareness among an in-market segment, then switching that same audience to CPC once enough data shows which shoppers are converting. The awareness phase establishes reach cheaply through viewable impressions, and the performance phase then captures the shoppers worth paying more for.
A third pattern pairs dynamic segments with a narrower purchases remarketing audience in the same ad group, comparing cost per acquisition between the two over a fixed test period. The dynamic segment often wins on reach, while the remarketing segment wins on conversion rate, and sellers who track both side by side can decide where to push budget rather than splitting it evenly by default.
Limitations and policies related to Sponsored Display audiences
Sponsored Display audiences come with a few practical limits worth knowing before you build a campaign. Not every audience type is available in every marketplace, and some pre-built Amazon Audiences or dynamic segments may not have enough matched shoppers to deliver meaningful volume for niche products.
Optimised targeting, while useful for scale, can stop serving a manual segment it judges inefficient, which means a campaign that relies entirely on automation may lose visibility into a smaller but valuable audience, according to Amazon’s optimised targeting documentation. Advertisers should treat this as a reason to audit regularly rather than a reason to avoid optimised targeting altogether.
AMC audiences require enough underlying data and signal volume to build reliable rule-based or lookalike segments, so very small sellers may not yet have the sales history to make full use of them. Self-service Sponsored Display has no minimum spend, which keeps testing accessible, but campaigns with very small daily budgets may struggle to gather enough impressions to judge an audience fairly within a normal reporting window.
Creative policies also apply: automated creative pulls directly from your existing listing content, so a weak product image or title will carry through to the ad, and custom image creative must still meet Amazon’s standard advertising content guidelines.
Integrating Sponsored Display audiences with Sponsored Brands and Sponsored Products
Sponsored Display audiences work best as part of a layered strategy rather than a standalone campaign. Sponsored Products and Sponsored Brands capture shoppers actively searching for relevant keywords, while Sponsored Display audiences reach people before or after that search moment, either building awareness ahead of it or remarketing to people who clicked but did not convert.
A practical sequence: use Sponsored Products to win the immediate search, Sponsored Brands to build brand presence across related search terms, and Sponsored Display remarketing to recover the shoppers who saw either of those ads but left without buying. This guide on how Sponsored Display works alongside Sponsored Brands covers the mechanics of that handoff in more detail.
Budgets should reflect this division of labour rather than being split evenly. Search-based formats typically deserve the larger share since they capture higher-intent demand, while Sponsored Display audiences work as a smaller, targeted layer that extends reach and recovers near-misses. A broader comparison of Sponsored Brands against Sponsored Products is useful context for deciding how much of the budget search formats should hold before audience campaigns get their share.
Where audience targeting fits in a full-funnel strategy
Sponsored Display audiences do their best work filling the gaps either side of search: building awareness before a shopper searches and recovering attention after they click away. I would pause audience spend and shift budget to Sponsored Products or Sponsored Brands the moment NTB share drops and CPA climbs past what search delivers. For smaller sellers, start with one remarketing segment before testing broader audiences.
— Harry
Getting more from your Sponsored Display audience tests with Osellpa
Running the tests this guide describes, different lookback windows, manual versus optimised segments, CPC versus vCPM, takes time if you are pulling Amazon reports by hand every week. The software pulls Sponsored Display data through Amazon’s API and turns it into profit-level reporting and PPC bid optimisation reports, so users can see which audience and bid combination is actually making money rather than just generating clicks.
Automated bid optimisation can sit alongside alerts that flag a spend spike or conversion drop before it eats a week’s budget, which matters most when running several audience variants side by side. Sellers can start with a free Amazon PPC bid optimisation report to see how their current Sponsored Display audiences are performing, or look at the Launch, Scale and Advanced plans to find a fit for ongoing campaign management.
FAQ
What are the four types of target audiences?
In Sponsored Display, the core audience types are Amazon Audiences, views remarketing, purchases remarketing and dynamic segments, each matching a different stage of the buying decision. AMC audiences add a fifth option for advertisers who want rule-based or lookalike segments beyond these pre-built groups.
Can I use sponsored display ads on Amazon?
Yes, Sponsored Display is a self-service advertising format available to sellers, vendors and advertisers directly through the Amazon Ads console, with no minimum campaign spend required. It works alongside Sponsored Products and Sponsored Brands as part of a full advertising strategy.
What are examples of display ads?
Sponsored Display examples include an automated ad using your existing product images shown on a competitor’s detail page, a custom image ad remarketing to shoppers who viewed your listing, and a vCPM awareness ad served to an Amazon Audience segment on a third-party website. Dynamic creative optimisation can also automatically pick the best-performing version when a campaign includes several products.
How much do display ads pay?
Sponsored Display uses either CPC billing, where you pay per click, or vCPM billing, where you pay per thousand viewable impressions, and the right choice depends on whether the campaign goal is conversions or reach. There is no fixed rate since bids are set by the advertiser and vary by audience type and competition.
How do I choose the right lookback window for my product?
Match the window to how long shoppers typically take to decide: around 7 days for fast-moving goods and 14 to 30 days or longer for considered purchases. Testing two windows side by side for a few weeks is the most reliable way to confirm which one fits your product.
