The Amazon demand-side platform (DSP) is the premier system for advertisers to connect with shoppers at scale, on almost any device, on and off of Amazon through programmatic ad purchasing.
With the Amazon DSP, advertisers can create impactful video and display ads and use Amazon’s unparalleled consumer purchase data to target highly qualified audiences at all stages of the customer purchase journey. All of this while automating purchases of online ad space from the Amazon demand-side platform’s extensive and exclusive ad inventory.
This compendium was designed to be your source of truth for everything related to Amazon DSP advertising. Whether you’re looking for explanations of what it is and how it works, details on specific inventory types, or tips and strategies on how to set up and optimize your campaigns, our compendium can help you along every step of your Amazon DSP journey.
Part 1: The fundamentals: Amazon DSP advertising
In part 1 of our guide to Amazon DSP, we run through the most important fundamentals, from what Amazon DSP is, to why you should use it and how much it costs.
What is the Amazon demand-side platform (DSP)?
The Amazon demand-side platform (DSP) allows advertisers to buy video and display ads at scale, both on and off of Amazon, using programmatic ad purchasing. Harnessing the power of shopper and advertising inventory data, programmatic ad purchasing automatically selects which digital ad slots to buy, and at what price.
So what can Amazon DSP advertising do for you? Here’s a quick summary of the most exciting aspects of the Amazon demand-side platform:
Amazon-exclusive insights and shopping signals to target relevant audiences
robust reporting to track progress towards your goals
availability for advertisers that don’t sell on Amazon and those that do
What is the difference between Amazon DSP and Sponsored Ads?
Sponsored Ads refers to Sponsored Products, Sponsored Brands, and Sponsored Display ads. Generally, each ad type is a self-service solution that targets a distinct area of the advertising funnel. Sponsored Ads appear primarily on Amazon, where shoppers go in order to consider a product or make a purchase.
The Amazon DSP, however, offers advertisers full-funnel advertising capabilities—both on and off of Amazon—with the option to choose between self-service and managed-service. Because the Amazon DSP has greater off-Amazon advertising capabilities than Sponsored Ads, it can reach shoppers earlier in the purchase journey, driving brand awareness and consideration.
The Amazon demand-side platform (DSP) and Sponsored Ads also differ in their eligibility requirements, targeting mechanisms, and cost structure. We’ve summarized all of the essential differences in the table below.
Why should I use the Amazon demand-side platform (DSP) for advertising?
In the first half of 2022, Perpetua conducted a DSP study that analyzed approximately $18 million in Amazon DSP advertising spend in the United States. Its findings highlight four core benefits that the Amazon demand side platform provides. We found that advertisers using the Amazon DSP see:
Benefit 1: improved Sponsored Products performance
Benefit 2: a larger share of mid- to long-term purchases
Benefit 3: a larger share of new-to-brand customers
Benefit 4: performance in line with, and sometimes surpassing, Sponsored Ads
Benefit 1: The Amazon DSP boosts the performance of Sponsored Products ads
Perpetua’s analysis found that Sponsored Products ads performed better when used in combination with the Amazon DSP. Conversion rate was 19% higher for Sponsored Products ads that were run alongside the Amazon DSP than for Sponsored Products ads used by themselves. Similarly, Sponsored Products ROAS was 3% higher when used alongside the Amazon DSP.
Benefit 2: Streaming TV (STV) video ads and the Amazon DSP see a large share of mid- to long-term purchases
Although they are a part of the Amazon demand-side platform, for this section of our analysis we placed Streaming TV ads (STV) into a separate segment to better understand what type of customers these brand-centric, large-screen-format ads reach.
We found that 70% of STV purchases occurred within 3-14 days after first ad interaction, followed by 33% for DSP purchases (excluding STV), and only 9% for Sponsored Products.
These results reflect that DSP and STV ads are more focused on driving awareness and new-to-brand sales than Sponsored Products, and that they are an effective way of building mid- to long-term demand.
Benefit 3: Streaming TV (STV) video ads see a greater share of new-to-brand customers than other Amazon ad types
Advertisers looking to introduce new customers to their brand and product portfolio often use the Amazon DSP to run Streaming TV video ads (STV, also known as OTT).
Our study showed that the share of new-to-brand purchasers was higher among those customers who had seen an STV ad than those who had seen only other types of advertisements on Amazon. Among customers who were influenced by an STV ad, 57% were new-to-brand versus only 45% among customers who were not exposed to STV.
These results suggest that advertisers who use STV ads with the Amazon DSP are more likely to be successful in gaining new-to-brand customers than those that do not use STV.
Benefit 4: Amazon DSP advertising is high-performing
We found that the Amazon demand-side platform’s ROAS was 43% higher than Sponsored Display, 9% higher than Sponsored Brands, and only slightly lower (-4%) than Sponsored Products ads. Note that for our ROAS analysis, we used the standard attribution models for each ad format—”last-click” conversion for Sponsored Products and “view-through” conversion for the Amazon DSP.
Part 2: Amazon demand-side platform (DSP) inventory, targeting, and creative types
The Amazon DSP provides advertisers with a variety of different creative, targeting, and inventory types to help them achieve their advertising goals. Because a wealth of choices can make for difficult decisions, we’ll go through all of them below, explaining key features and use cases.
Amazon DSP advertising inventory
Before getting into the different types of Amazon demand-side platform advertising inventory, let’s take a look at what the term ad inventory means, and how it is different from a supply source.
What is Amazon DSP ad inventory? And how is it different from a supply source?
Ad inventory refers to the channel where you want to display your ads. An example of an ad inventory type would be Amazon Owned and Operated mobile (AZOO MB). A supply source is the specific exchange or site your ads will show on. So for example, Amazon mobile app, Amazon mobile web, and IMDB mobile web are three examples of AZOO MB supply sources.
Thinking of inventory types as one level above supply sources can help conceptualize the relationship.
Types of Amazon demand-side platform inventory
The Amazon DSP’s advertising inventory allows you to reach Amazon audiences at scale wherever they are: on a desktop or laptop, phone, or streaming device. The Amazon DSP offers the following inventory types:
Amazon owned & operated mobile (AZOO MB), desktop (AZOO DS), and mobile app (Mobile AAP): exclusive inventory from Amazon retail sites and O&O properties such as IMDb
Amazon Publisher Services (AZPS): inventory from leading publishers via Amazon Publisher Service’s header bidding integrations
Third-party exchanges (3P): inventory on sites not owned or operated by Amazon, but available through leading exchanges such as Google Authorized Buyers
When setting up your Amazon DSP campaigns, we recommend using a variety of inventory types. This provides scale to achieve your campaign goals. It also diversifies ad formats and content for all customer touchpoints. And finally, by utilizing different supply sources, you can see which supply source is performing best and optimize your campaign accordingly.
To give you an idea of the performance of each supply source, we’ve provided H1 2022 benchmarks from our Amazon DSP Deep-Dive. For more, read the report.
Amazon DSP Metric
Total cost per thousand impressions
Click-through rate (total clicks / total impressions)
Cost per click
The number of detail page view conversions relative to ad impressions (DPVR = DPV/Impressions)
The average cost to acquire a detail page view
Purchase rate (purchases / impressions)
New-to-brand purchase rate (new-to-brand purchases / impressions)
Return on ad spend (ad revenue / ad spend)
Average order value (ad revenue / number of orders)
Using the Amazon demand-side platform, advertisers can locate, engage with, and activate shoppers at all stages of the purchase journey. Each targeting type maps to a different stage of the advertising funnel.
Below is a table that describes each targeting type and what it does.
The Amazon demand side platform offers ad creatives that can be used to target any area of the advertising funnel. Generally, video ads are most effective at driving awareness and consideration, while display ads create purchase intent and turn shoppers into customers.
Video ad types
The two main video ad types offered via the Amazon DSP are Streaming TV ads (STV) and online video ads (OLV).
Streaming TV ads (STV, also known as OTT)
Streaming TV (STV) ads play before and during content on streaming platforms. Similar to traditional television commercials, they cannot be skipped and they cannot be clicked. Services like IMDb TV, Fire TV, and 3rd party on-demand apps are where STV ads appear. Here are some best practices to keep in mind when creating STV ads:
A logo on the creative throughout the ad can help increase brand recall and relevance.
Selecting the CTA carefully can ensure it’s aligned with the campaign objective.
Emotional immediacy and original messaging can help to capture your audience’s attention and create a memorable brand experience.
Online video ads (OLV)
Online video ads (OLV) ads appear on web browsers and apps before or during video content. These ads are clickable and can lead back to Amazon product detail pages or to your own website. The following best practices can help you craft highly-effective OLV ads:
Introduce your brand early in the video, especially if you’re including speech. A study by Amazon showed that detail page view rate (DPVR) fell by 22% when speech was delayed by even one second in an ad.
Avoid micro-videos. The same study showed that videos that lasted between 24 and 29 seconds had a DPVR 181% higher than video ads that were 6 seconds or less.
Use consistent voices. The optimal number of speakers is 2, increasing DPVR by 10% compared to just 1 speaker. More than 2 speakers may decrease DPVR by up to 45%.
You can find more video ad best practices for the Amazon DSP here.
Now we can look at the various options for display ad creatives available to Amazon DSP advertisers.
Responsive eCommerce creative (REC)
Responsive eCommerce creatives (REC) have the same features as dynamic eCommerce ads. However, RECs can generate an ad for up to 20 ASINs per creative—and they can be served on all five Amazon DSP inventory sources. Amazon provides reporting across all creative sizes so advertisers can determine which placement performs best for each creative.
The table below outlines many of the key differences between DEAs and RECs.
More details about Amazon’s eCommerce display creatives can be found here.
Custom image creative (Custom Image)
Custom image creative (custom image) uses brand-specific design, layout, and copy, replacing the product detail page image. This ad type links to product detail pages, store pages, custom landing pages, or Whole Foods Market.
Amazon Desktop, Amazon mobile, and third-party websites are the inventories where custom image ads can be served. This creative type is ideal for an advertiser who is looking to craft a more brand-centric experience because it provides the opportunity to use brand-specific messaging, brand images, and branded fonts.
Custom dynamic eCommerce creative (DEA Custom Image)
A custom image can be added to eCommerce creatives to include brand imagery, branded fonts, and custom copy. Images are dynamically inserted into ad units of various sizes. Again, the goal of using DEA Custom Image is to serve an ad that captures some of your brand’s unique look and feel.
Display ads best practices
For display advertising campaigns, focus on clarity and getting your brand message across concisely. Here are some guidelines to craft ad creative that will appeal to shoppers:
Imagery should be clear, either product-focused or lifestyle; it’s best to keep the product as identifiable as possible.
The logo should be visible to identify your brand and encourage recall and affinity.
Copy should be simple and direct, focusing on the unique features of the product.
Headlines should communicate a strong statement that provokes action.
For more information, visit the Amazon DSP Help Center where you can find additional best practices and strategic advice.
Performance benchmarks by creative type (H1 2022)
Part 3: Goals, campaign setup, evaluation, and optimization
Now that you’ve got a handle on what the Amazon demand-side platform is, and the range of options it offers, it’s time to learn how to get started. This section of our compendium looks at setting goals, building campaigns, and evaluating and optimizing your Amazon DSP ad performance.
Setting goals for your Amazon demand side platform campaign
Before setting up your campaign, it’s essential to have a clear goal. This is especially true because you can use the Amazon DSP to target any area of the funnel or multiple areas at once—and because certain creative, inventory, and targeting types are better suited to some goals than others. To keep things simple, we’ll outline three goals that can guide your decision-making as you set up your Amazon DSP advertising campaign:
Brand awareness: The objective of brand awareness is to introduce people to your company and its products, and to ensure that your brand stays top of mind throughout the purchase journey.
Consideration: A consideration campaign aims to address a pain point, highlight an interest, or answer a question for consumers so that you, the advertiser, can show how your product meets the shopper’s need.
Conversion: This is the end goal for every marketer—sales. You’ve introduced yourself and educated the audience about your offer. Now it’s time to stand out, differentiate your products from the competition, and encourage a purchase.
We’ll pick up with more tips on how to structure and optimize your campaigns for the goals we’ve outlined above in the next section.
In this section, we cover all the key settings you need to understand when creating your Amazon DSP campaigns. We also talk about what “self-service” and “managed-service” mean in practice.
Managed-service vs. self-service
Before getting into the finer points around campaign setup, advertisers interested in using the Amazon DSP need to understand that they have two potential options: self-service and managed-service.
With the managed-service option, someone else will take care of the campaign strategy, execution, and optimization. This is the ideal solution for an advertiser that lacks the know-how or time to develop and implement their own advertising strategy.
In contrast, an advertiser interested in a self-service solution would need the time and know-how to implement and fine-tune their own campaigns on an ongoing basis.
Understanding the views in the DSP console
The Amazon DSP consists of a hierarchy of different levels. From the Amazon DSP campaign manager page, check the top left-hand side of the screen to identify which level of the hierarchy you’re on. Your level will be accompanied by information about your selection (advertiser name, order name, line item name, creative name, etc.).
Order setup best practices
Once you have established yourself as an advertiser within your new entity in the Amazon DSP console, the first step to setting up a new campaign is to create an order. Remember that you cannot delete orders once they are created, so consider the following details carefully before proceeding.
Order naming convention
Following a uniform naming convention for your orders will make navigating them and analyzing their performance easier. We recommend including the product name, the targeting tactic, the goal KPI, and the level of the funnel in your order name. This way, looking at all the orders in the advertiser view, you can quickly identify an order’s advertised product, its goal, and the success criterion.
Our recommended best practice for naming orders is as follows:
Selecting an order goal and KPI
Next, you’ll need to select the goal for your order and the KPI that best aligns with the creative you plan to use. Below is a table to help you decide which KPI to choose. Note that Streaming TV (STV) ads are non-clickable and non-skippable, so any video metrics that gauge viewer clicks or completions are unsuitable for STV ads.
For your bid strategy, we recommend choosing “While spending full budget, maximize performance.” We’ve found that campaigns with this setting pace better—i.e., spend more consistently and evenly—than those that use the “maximize performance” setting.
Max bid is the primary bid control when using the “while spending full budget, maximize performance” bid strategy, and “base bid” is the primary control when using the “maximize performance” bid strategy.
Budget and delivery
Flights allow you to further segment your budget by date. Up to twelve flights can be created at a time. Our recommended best practice is to create one flight per month.
The budget cap option limits the order ad spend by day or month, allowing you to further calibrate your budget. We recommend setting a monthly budget cap, which ensures that your full budget will be spent each month and that your campaign will run with optimal pacing.
Line item setup best practices
After you’ve created your order, you’re ready to set up line items. Each line item type provides access to different devices, inventory types, supply sources, and targeting options.
Our recommended best practice is to create one line item for each of the five inventory types. Having line items segmented by supply source will make it easier to identify which is performing best, allowing you to allocate resources accordingly.
Line item settings
Type: Choose the type of creative you would like to run (display, video, etc.).
Name: In your line item name, include the product being advertised, the targeting tactic, and the inventory type. Our best practice naming convention goes as follows: Product_Tactic_InventoryType. Example: PlayStation5_Retargeting_AZOO_DS.
Category: Select the category that aligns best with the advertised product. For maximum reach, you can choose the whole category that encompasses your advertised product. For example, if you’re advertising lipstick and want to maximize reach, you would select the “Beauty & Fashion” category.
Inventory: Choose the inventory the line item will run on. By default, the line item will run on all available real-time bidding exchanges, Amazon-owned sites, and Amazon Publisher Services (APS) inventory.
A recommended best practice is to separate line items by Amazon-owned sites, APS, and 3rd party bidding exchanges to have more control over bucketed bidding strategies.
Because the selected inventory limits your line item and creative type choices, you should know where you’d like your ads to serve or the supply source before setting up your campaigns.
Below are two tables that map inventory types to device type, ad environment, line item type, and supply source for display and video ads. You can use this for reference when creating your campaigns.
Targeting: This is where you can select the device type, the mobile environment (web vs. app), region, and more.
Audiences: Audiences allow you to reach shoppers in a highly-specific manner in the Amazon DSP. Amazon Ads offers two types of Amazon audiences based on shopping interactions:
Pre-built Amazon audiences based on Amazon shopping and streaming signals
Custom Amazon audiences based on shopping interactions with your brand’s products or related products in Amazon’s store
Below, we go through a quick rundown of the different types of audiences you can target.
Custom audiences based on your advertising data
Product category audiences that reach Amazon customers when viewing products within specific categories in Amazon’s store
Custom audiences based on interactions with your brand’s products or related products
Audiences based on age, gender, income level, education, geography, and household size
Users visiting from specific devices or browsers
Groups of customers that have been browsing for products within a certain category in the last 30 days
Audiences based on what shoppers frequently browse and buy
Audiences shopping for products related to specific milestones in life, like moving, getting a pet, or having a child
Based on shopping and viewing activities that suggest certain lifestyles, like gamers, movie enthusiasts, foodies, etc.
Shoppers that have similar purchase and browsing histories to existing customers
Audiences based on off-Amazon activity that can provide a more holistic view of all the relevant audiences, both on and off of Amazon
Campaign setup best practices
Run cross-device campaigns to find shoppers wherever they are—on sites, apps, and devices. This will extend your reach and ensure your brand is always top of mind.
Segment line items. At the beginning of this section, we recommended bucketing your supply sources. Depending on the number of line items you’re creating, it may also make sense to break out your campaign’s line items by device and audience type. This added layer of segmentation will help you achieve more granular performance measurement, and it will allow you to optimize more easily.
Build out audiences on both Amazon O&O inventory and across third-party sites and apps. Incorporating diverse inventory and supply sources into your campaign increases the likelihood of reaching relevant audiences.
Analyzing Amazon DSP campaign performance
Before fine-tuning your campaigns, you first need to understand what is working and what isn’t. Collecting Amazon DSP advertising performance data takes time because of the platform’s 14-day attribution window. Below, we explain the attribution model, plus important KPIs and reports that will help you measure performance.
The Amazon demand-side platform attribution model
Attribution models determine which ad should get credit for a conversion event. In the Amazon DSP, conversion events include purchases and other types of events, like viewing a detail page or adding an item to a cart.
The attribution model also determines the length of the “look-back window,” or the timespan within which an ad interaction can happen in order for it to be credited for a conversion event. This is why understanding how the Amazon DSP attribution model works is essential to interpreting the metrics in the Amazon DSP’s reporting.
For display and video ads, the Amazon DSP uses a 14-day, “last-touch” attribution model. This means the conversion will be credited to the last customer touchpoint with an ad—that is, a view or a click—that took place up to 14 days before the conversion event.
The exception to this model is that clicks trump views. The last click that took place within the 14-day lookback window will receive credit, even if there are views that occur later.
For more information on the Amazon demand-side platform’s attribution model, check out this article.
Amazon demand-side platform (DSP) performance metrics
The Amazon DSP provides advertisers with metrics to track their performance at every stage of the customer purchase journey. Below is an overview of the key metrics for evaluating your progress toward your goals.
When analyzing the performance of your campaigns, it’s important to remember that results take time. Because of the Amazon DSP’s 14-day attribution window, there may be a delay of up to 14 days in learning that one of your ads has been credited with a sale or another conversion event.
This is just an overview of some of the metrics that the DSP reporting provides. This article provides a more extensive list of metric definitions and use cases.
Reporting in the Amazon DSP
The Amazon DSP provides advertisers with access to comprehensive and in-depth reporting that will help you monitor and make decisions related to your campaign goals. Below is an overview of the key reports and the information they provide.
The console reporting provides you with all of the fundamental performance information about your Amazon DSP advertising campaigns. From here, you can evaluate your campaign data at the order, line item, and creative level, with more than 100 metrics to choose from.
Using audience insights, advertisers can view highly granular data on customer purchases and views. You can, for example, see the time elapsed between when a customer views a product page and makes a purchase (pictured above). View and purchase data can also be segmented by demographic status (e.g., age, gender, household income), timing (hour of day, day of week, and weeks/months of the year). Data on purchase frequency—like repeat purchase rate, number of repeat purchases, and time between repeat purchases—is also available via audience insights.
Overlap reports allow you to discover other audiences your target audience is a part of. This is a great way to identify opportunities for increasing your reach and introducing new shoppers to your brand and product portfolio. The affinity score measures the similarities between a target and an overlapping audience. The higher the affinity score, the more similar the audiences are. Size measures how large the audience is compared to all Amazon audiences. A size of 6, for example, would indicate that the audience is larger than 60% of all Amazon audiences. Taken together, size and affinity scores help you identify the most significant complementary audiences that may be interested in your products, opening up opportunities to attract new customers to your brand.
The segmentation report allows you to evaluate the performance of directly targeted segments and uncover new segments you have not yet targeted. This report can be a great way to identify high-potential audience segments you have not yet targeted. Using the example above, you can easily sort by click-through rate (CTR) to see that the 50+ Active and Healthy Living Products line item has a CTR of 12.6%.
This is a great start, but you can also use a pivot table to group the line items according to segment, and analyze performance at the segment level. In our example, we see that the segment with the best DPVR overall is 50+ Active and Healthy Living Products, with a detail page view rate (DPVR) of 26.1%, making this an ideal audience to target.
Retail insights report
The retail insights report can help you understand shopping activity for your brand’s products on Amazon before, during, and after a display or video advertising campaign. This allows you to gauge whether your video ads are having the intended impact. Metrics in this report include: search impressions, earned media impressions, glance views, orders shipped, conversion rates, number of reviews, average ratings, and percentage of 1- and 2-star reviews.
Optimizing your Amazon demand-side platform (DSP) campaigns
Once you’ve understood how your campaigns are performing and identified the areas in which you’d like to see improvement, you can begin to make changes to ensure that your campaigns achieve their intended goals.
In this section, we’ll review the different manual levers and strategies that advertisers can use to optimize their campaigns. Most advertisers will either optimize towards delivery (that is, how often the ad is served) and/or performance (frequency or cost of conversion).
Levers for manual optimization
Audiences: Advertisers should make sure the audience that is being reached is right for the product the campaign is promoting.
Before campaign setup, advertisers may reference overlap reports. These reports can be built from two sources:
from the advertiser’s audience through the use of a pixel traffic, DMP, or advertiser hashed audience
from Amazon audiences which are created based on shopping interactions and built through audience builder
Once the campaign is live, advertisers can download an audience report to see the performance on selected and unselected audiences.
Alternatively, advertisers can use historical campaign reports to identify top-performing selected and unselected audiences to target.
Bids: How much an advertiser is willing to pay for an impression, represented in cost per thousand (CPM).
Bid strategy: Set up base and maximum bids, then shift bids and allocate more budget to high-performing line items. Use incremental bid increases to increase scale while avoiding overpaying for inventory. The right base bid and maximum bid for a line item depends on factors like optimization type, audience segments, supply sources and budget.
Budget: Shift from under-delivering lines into highest-performing lines that are also scaling.
Creatives: Optimize to show the highest-performing creatives and remove underperforming ones.
Frequency: The number of ad impressions seen within a given time period.
Frequency cap: A set limit on the number of times a unique user can see an ad during a specific time period.
When to adjust and why: Increase the frequency cap to serve more impressions to a line item meeting or beating the campaign goal, yet under-delivering (i.e., not displaying enough).
Supply: Select any unchecked supply sources if you notice that a campaign is under-delivering.
Viewability: Measurement of how viewable an impression is based on the visibility of an ad when it was served. The MRC (Media Rating Council) defines viewability as when 50% of pixels are in view for at least 1 second.
Viewability tiers: greater percentage = greater viewability
When to adjust and why: relax the viewability setting to serve more impressions to a line item that is performing well yet under-delivering.
The table below outlines key best practices for Amazon DSP advertising optimization.
The Amazon DSP is a massive opportunity for online advertisers, offering ad targeting based on Amazon’s unparalleled consumer purchase data, exclusive and comprehensive ad inventory, and industry-leading campaign reporting and insights.
A couple of key takeaways:
Apart from having high-performing ads, advertising with the Amazon DSP also has ancillary benefits like improved Sponsored Products performance, plus a higher share of mid- to long-term purchases and new-to-brand customers;
Amazon DSP advertisements can target any area of the funnel, yet STV and OLV ads can provide unique upper-funnel reach to build your brand.
Remember that the Amazon DSP has a 14-day attribution period and that results may take time to appear in your reporting and your dashboards.
Solutions like Perpetua’s DSP optimization software can simplify an inherently complex platform, allowing you to quickly build orders, generate line items, and efficiently scale ROAS. Get started with Perpetua’s Amazon DSP ad engine.