Sponsored ads target words. Amazon DSP targets people. That single difference is why DSP can reach shoppers who never searched for your product, and why a DSP campaign with the wrong audience spends money with nothing to show for it. The targeting options are the whole game.
This guide lays out the audience families, what each one is for, how they combine, and the order a seller new to DSP should try them.
The audience families
Retargeting
Audiences built from what shoppers did with specific products. Viewed your product and did not buy. Bought your product and might buy again. Viewed a competing product. Added to cart and left. These are the smallest audiences and the ones with the most intent, because every member has already interacted with something relevant.
Retargeting is where DSP earns its first dollar for most brands. The audience is defined by your own catalog, the window is adjustable (viewed in the last 7, 14, 30 days and so on), and the result is measurable against a clear baseline.
In-market
Shoppers whose recent behavior across Amazon suggests they are actively shopping in a category right now. Amazon builds these segments from browsing and purchase signals and names them by category. A seller of kitchen knives can target the in-market segment for cutlery, or for cookware more broadly, or for kitchen and dining as a whole.
In-market audiences are larger than retargeting and closer to purchase than lifestyle. They are the usual second step: once retargeting works, in-market is how the campaign reaches shoppers who have not seen your brand but are buying in your space this week.
Lifestyle
Broader segments built from long-run purchase patterns. Outdoor enthusiasts. New parents. Home cooks. Premium beauty shoppers. These describe a type of person rather than a current shopping trip, and they are much larger and much less immediate than in-market.
Lifestyle audiences are an awareness tool. They put the brand in front of the right kind of shopper at a moment when that shopper is not necessarily buying. The return shows up over weeks as lift in branded search and in-market conversion rather than as direct attribution, and campaigns run on lifestyle audiences alone are difficult to read.
Contextual
Placement based on the content the ad appears next to rather than on the shopper. A product page in your category. A piece of content on a publisher site about your product type. Contextual targeting does not use shopper data and is useful for reaching people where the context does the targeting for you.
Lookalike and custom
Lookalike audiences take an existing audience, usually your buyers, and find shoppers whose behavior resembles them. Custom audiences let advertisers build segments from combinations of signals or from their own uploaded data where Amazon permits it. Both are tools for scaling what already works, not for finding what works in the first place.
Demographic and device
Age ranges, household signals, device type, and similar. Mostly used as filters on top of another audience rather than as the targeting itself. A lifestyle audience narrowed by a demographic filter is a common shape.
How layering works
Audiences can be combined with "and" logic (must be in both), "or" logic (in either), and exclusions (in this one but not that one). The useful patterns:
Retargeting with a purchase exclusion. Viewed your product, did not buy it. The single most common DSP line item.
In-market narrowed by a lifestyle filter. Shopping for cookware now, and a home cook by pattern. Smaller than either alone, more relevant than either alone.
Competitor viewers with your buyers excluded. Looked at a competing product, has never bought yours.
Every layer makes the audience smaller and the ad more relevant, up to the point where the audience is too small to deliver. The skill in DSP is finding that point, and it moves by category and by catalog size.
The order to try them
One. Retargeting your own product viewers, purchase excluded, 30-day window. Run it long enough to read, which for most catalogs is at least two to four weeks. This tells you whether DSP creative and landing pages work for your products at all, on the audience most likely to say yes.
Two. Retargeting past buyers for products with a repurchase cycle. Consumables, refills, anything bought on a schedule. Window set to the typical cycle.
Three. In-market for your closest category. This is the first audience that reaches new shoppers. Compare its cost and conversion to the retargeting line directly.
Four. Competitor viewers, if the category allows it and the product has a clear advantage to show.
Five. Lifestyle and lookalike, once the campaign has enough history that lift in branded search and in-market performance can be measured against a baseline.
Adding steps three through five before one and two has a predictable result: large spend, unclear attribution, and no way to tell whether the problem was the audience, the creative, or the product page.
What stays constant
DSP audiences decay. A retargeting audience from a Q4 promotion is not the same audience in February. In-market segments shift with season. Audiences need review on a schedule the same way sponsored ad campaigns do, and DSP budgets need the same event ramps and restores. The habits that make sponsored ads monitoring work carry over directly, at a larger budget. This comparison covers where DSP sits relative to the sponsored ad types.
Frequently asked questions
What is the difference between in-market and lifestyle audiences in Amazon DSP?
In-market audiences are built from recent shopping behavior that suggests someone is buying in a category now. Lifestyle audiences are built from longer-run purchase patterns that describe a type of shopper. In-market is closer to purchase and usually converts better; lifestyle is broader and better for awareness.
Can Amazon DSP retarget shoppers who viewed a competitor's product?
Yes. DSP can build audiences from views and purchases of specific ASINs, including ones you do not sell, within the rules Amazon sets for competitor targeting. This is one of the most common uses of the platform and one of the hardest to do well, because the shopper has already shown a preference.
How many audiences should a new DSP campaign target?
Start narrow. One retargeting audience of your own product viewers, run long enough to read, before adding in-market or lifestyle segments. Layering several audiences at once makes it impossible to tell which one is producing the result.
Off Hours runs the sponsored ads foundation that DSP is layered on: dayparting, budget, event and performance rules on a 15-minute cadence. Start a free 14-day trial.