Key Takeaways
- Amazon DSP gives sellers direct access to audiences who recently purchased from competitors, using purchase timing data to intercept buyers before they reorder.
- Product Views audiences let you target shoppers who visited a rival’s listing but did not buy, capturing undecided buyers with a clear switching reason.
- Complementary product targeting reaches buyers who already spend in your category but have never been exposed to your brand.
- Lookalike audiences built through Amazon Marketing Cloud find new buyers who shop the same way as your best existing customers.
- The DSP overlap report surfaces high-affinity audience groups that algorithms identify automatically, often outperforming audiences selected manually.
- All five strategies are available to brands running DSP campaigns, and each targets a different stage of the buyer decision process.
General Summary
Amazon DSP gives e-commerce brands access to five distinct targeting strategies for reaching competitor audiences: purchase timing retargeting, product views targeting, complementary product audiences, lookalike modeling through Amazon Marketing Cloud, and algorithm-driven overlap audiences. These strategies exist because Amazon holds purchase, browsing, and subscription data across its entire marketplace, data no third-party platform can replicate. Brands spending between $30,000 and $550,000 per month on Amazon advertising use these methods to reach shoppers at the exact moment a purchase decision is forming. The strategies differ in who they target and when: some go after buyers about to reorder, others target shoppers still comparing options, and others find buyers who have never encountered the brand. Each approach works best when paired with the right creative and the right exclusion logic.
Extractive Summary
Amazon DSP lets sellers build audiences from competitor product ASINs, targeting buyers in the window just before they reorder. Product Views audiences capture shoppers who visited a rival’s listing without buying, identifying people who have not yet committed to a brand. Complementary product targeting reaches buyers whose habits already resemble those of your best customers, without targeting rivals directly. Lookalike audiences use your existing subscribers and repeat buyers as a seed to find new shoppers with matching patterns. The DSP overlap report surfaces audience groups Amazon identifies automatically based on behavioral similarity to your buyer base.
Abstractive Summary
Amazon’s advertising infrastructure has turned purchase behavior into a targetable asset. Where traditional digital advertising relies on interest signals or demographic proxies, DSP uses actual transaction data: what someone bought, when they bought it, and what they tend to buy next. For brands competing in crowded categories, this creates a structural advantage. The question is not whether a rival’s customer can be reached. The question is which of the five available methods fits the moment, the product cycle, and the budget. Brands that treat DSP as a competitor intelligence layer, rather than a standard ad channel, see consistently stronger returns than those running it as a reach extension tool.
What Is Amazon DSP and Why Does It Target Competitor Audiences?
Amazon DSP (Demand-Side Platform) is a programmatic advertising platform that lets brands buy display and video ad placements using Amazon’s first-party shopper data. Unlike Sponsored Products, which targets keyword searches on Amazon, DSP targets specific audience segments based on past purchase behavior, product views, and subscription patterns. This distinction matters because DSP can reach shoppers off Amazon, on third-party sites and apps, while still drawing on Amazon’s purchase data to define who sees the ad.
Amazon holds purchase records for hundreds of millions of buyers. It knows what they bought, how often they buy it, and when they are likely to reorder. No advertising platform outside of Amazon can access this data. For sellers competing in categories where rivals hold a large share of repeat buyers, DSP is the primary tool for entering that buyer relationship before the next order locks in.
The five targeting methods below each use this data differently. Some go after buyers right before a reorder. Others capture shoppers who are still comparing. One goes after buyers who have never searched your category at all but already behave like your best customers. Each one targets a different point in the decision process.
How Do You Catch Competitor Buyers Right Before They Reorder?
You build a Product Purchases audience using your rival’s ASIN, then set the time window to match the product’s natural cycle. If someone bought a 30-day probiotic from a competitor 25 days ago, their bottle is nearly empty and the reorder decision is forming. DSP lets you find those buyers and place your ad in front of them at that exact moment.
In DSP, open the audience builder and select Product Purchases. Enter the rival’s product ASIN. Then set the lookback window to match consumption: 25 to 35 days for a 30-day supply, 80 to 95 days for a 90-day supply. The window is not arbitrary. It reflects the product’s own cycle and catches buyers when the choice is still open.
Two exclusion groups are essential here. Remove anyone currently on your own Subscribe and Save, because they are already committed to your brand. Also exclude anyone who purchased from you in the last 60 days. What remains is a list of shoppers who bought from your rival, are running low, and have not yet placed a new order.
The creative for this audience does not need to be a general brand message. Specificity converts better. A stronger dose, a better price per serving, faster delivery, or a higher review count: one clear reason to switch is enough. The buyer is already in decision mode. Make the decision easier.
How Do You Reach Shoppers Who Looked at a Rival’s Listing but Did Not Buy?
A Product Views audience targets shoppers who visited a competitor’s listing without completing a purchase. This is a fundamentally different buyer than someone who bought from a rival. They have not committed. They searched, they clicked, they left. That signals a comparison in progress, not a loyalty relationship.
The setup mirrors the Product Purchases audience. In the DSP audience builder, select Product Views and enter the rival’s ASIN. Set the lookback window at 14 to 30 days. Past 30 days, the signal degrades: shoppers have either bought from someone or moved on entirely. Inside that window, intent is still warm.
These buyers know the category. They understand what they are looking for. They just did not pick your rival. The ad needs to give them a single clear reason to pick you instead. Better reviews, more servings per dollar, a cleaner ingredient list, a subscription discount: one reason, stated plainly. They were already comparing when they left the rival’s page. Your ad is the next comparison point.
Exclude your existing buyers and subscribers from this audience as well. Every DSP campaign benefits from clean exclusion logic. It reduces wasted spend and keeps bid efficiency high.
How Does Targeting Complementary Products Build a Competitor Audience Without Going Head-to-Head?
Complementary product targeting reaches buyers whose purchasing habits already resemble those of your best customers, without targeting your rivals directly. If you sell a probiotic, a large portion of your buyers also purchase vitamin D, magnesium, or a greens powder. DSP lets you target people who buy or subscribe to those adjacent products.
The logic here is behavioral rather than competitive. A buyer who has maintained a magnesium subscription for six months is a proven repeat purchaser. They spend on their health consistently. They know what a quality product looks like. They are warmer than someone who searched for a probiotic once and bounced, because their buying behavior demonstrates a pattern your product fits.
Build this audience using Product Purchases or Subscribe and Save, but substitute a complementary ASIN instead of a rival’s. A 60 to 90-day window works well for supplements and consumables. Exclude your own buyers and subscribers as standard.
This method expands the addressable audience beyond people who are already in your direct category. It also sidesteps the bidding competition that builds up around high-volume rival ASINs. The buyers reached through complementary targeting often represent lower acquisition costs with comparable conversion intent.
What Are Lookalike Audiences and How Do You Build One From Your Best Buyers?
A lookalike audience uses your best existing customers as a data seed, then instructs Amazon to find other shoppers on the platform who behave the same way. Amazon matches shopping patterns: purchase frequency, category spend, subscription behavior, and order history. Many of the people who match that pattern are currently buying from your rivals.
The seed quality determines the output quality. Build the seed list through Amazon Marketing Cloud. Start by pulling your subscribers and repeat buyers. Then remove one-time purchasers and anyone who bought only during a heavy discount period. What you want in the seed is buyers who came in at normal price, stayed, and kept ordering. Those are the people whose pattern is worth replicating.
Amazon offers five match levels, ranging from a tight match to a wide match. Tight matches produce smaller audiences with higher behavioral similarity. Wide matches produce larger audiences with more noise. Start tight, test conversion rates, and open the match level as data accumulates.
Brands running AMC lookalike campaigns report returns of 3 to 5 times the performance of standard in-market campaigns. One case in the supplement category saw a 480% return increase over three months after switching from broad in-market targeting to AMC-seeded lookalikes. The difference is the seed: instead of targeting a category, you are targeting a behavioral profile built from real customer data.
How Does the DSP Overlap Report Find Audiences You Would Never Have Picked Yourself?
The overlap report inside DSP analyzes your existing audience and surfaces other audience groups with statistically high behavioral similarity. You are not selecting these groups. Amazon’s data is selecting them based on patterns across millions of transactions that no individual seller account can observe.
Start by building a base audience from your best buyers: subscribers, repeat purchasers, high-spend customers. Run the overlap report against that audience. Amazon returns a list of audience groups ranked by affinity score. The top groups share behavioral characteristics with your buyers in ways that keyword research and manual targeting cannot surface.
High affinity scores in the overlap report consistently outperform manually selected in-market audiences in conversion testing. The audiences identified tend to include buyers who have never purchased from you and may not have actively searched your category. They are reachable because they shop the way your customers shop, even if they are not searching for what you sell yet.
Take the top five to ten groups from the report and build separate campaigns for each. Allocate equal test budgets initially. After two to three weeks, measure cost per acquisition and return on ad spend for each group. Move budget toward the groups that convert. Cut the ones that do not. The overlap report does not produce a final campaign plan. It produces a set of data-backed hypotheses. Testing turns hypotheses into budget decisions.
Which of the Five Strategies Should You Run First?
Start with the purchase timing retargeting method if your product has a predictable consumption cycle. The ASIN targeting is straightforward to build, the audience intent is high, and the conversion signal is easy to interpret. This method works for any consumable category: supplements, pet food, coffee, cleaning products, personal care.
Run Product Views targeting alongside it from the start. The setup takes under an hour and the audience is low-cost to build. These are buyers who already know your category and have shown intent. They are the lowest-friction audience available through DSP.
Add complementary product targeting in the second month. Identify two or three products that your best customers commonly purchase together with yours. Build audiences from those ASINs. Test them against the direct competitor audiences on a cost-per-acquisition basis.
The lookalike and overlap report methods require a minimum customer base to seed effectively. If your Subscribe and Save list is under 500 buyers, the AMC seed will be too small to produce statistically reliable matches. Build the direct competitor audiences first, grow the customer base, and introduce AMC lookalikes when the seed has enough volume to generate a clean pattern.
All five strategies run inside DSP. The choice of which to run first depends on where your buyers are in the decision cycle and how much customer data you already hold. The reorder window and product views methods work with zero customer history. The lookalike and overlap methods pay larger returns as your own buyer data accumulates.
What Creative Do DSP Competitor Audiences Actually Need?
DSP creative for competitor audiences follows a different logic than brand awareness advertising. These are not cold audiences. They have already engaged with your category, in some cases with a specific rival product. The creative does not need to introduce the problem. It needs to offer a reason to choose differently.
One clear differentiator outperforms multiple claims in every test. Better reviews, lower cost per serving, faster delivery, a cleaner formula, a stronger money-back guarantee: pick the one that is most defensible and most visible in your listing. State it directly in the ad. The buyer is already comparing. Give them the comparison point.
For reorder timing audiences, urgency framing works. The buyer knows the product is running out. An ad that acknowledges the reorder moment, with a subscription discount or a free trial offer, converts better than a generic brand image.
For Product Views audiences, the creative should reflect what differentiated your listing when they were on your rival’s page. If your review count is significantly higher, lead with that. If your price per unit is lower, show the comparison. These buyers left a listing without buying. They were looking for a reason to pick one. Your creative is the reason.
How Do You Measure Whether DSP Competitor Targeting Is Working?
DSP reports cost per acquisition, return on ad spend, and new-to-brand percentage at the campaign and line-item level. New-to-brand percentage is the most important signal for competitor targeting. It measures how many purchases came from buyers who had not purchased your brand on Amazon in the past 12 months.
A strong competitor targeting campaign produces a new-to-brand percentage above 60%. If most conversions are coming from existing buyers, the audience exclusions are not working correctly or the campaign is overlapping with your existing customer base.
Compare return on ad spend across the five audience types. Purchase timing and Product Views audiences typically produce higher ROAS in early testing because intent is highest. Complementary product and lookalike audiences typically produce higher new-to-brand percentages because they reach buyers with no prior exposure to your brand.
Run each audience type as a separate line item inside DSP. This allows individual performance tracking without mixing signals. After four weeks of data, move budget toward the audiences producing the strongest combination of ROAS and new-to-brand rate. The overlap report audiences should be tested in the same structure: one line item per affinity group, equal starting budgets, four weeks of data before reallocation.
Amazon’s competitor targeting data does not stay static. Purchase patterns shift, rival ASINs change, and subscription rates fluctuate. Refresh audience definitions every 60 to 90 days to ensure the ASINs in your targeting still represent active, high-volume competitors. A rival ASIN with declining sales produces a shrinking audience with weaker intent signals. Keep the ASIN list current.
