Amazon Advertising Strategy 2020: What Still Works in 2026

Why "Amazon Advertising Strategy 2020" Is the Wrong Search Now

People search a year-stamped query when they are catching up, and most of the pages ranking for this one have not caught up either. The 2020 guides describe a world where Sponsored Products was the whole game, DSP was an enterprise line item, and "strategy" meant picking a new ad format to try. That world is gone.

What actually changed since 2020 is quieter and less exciting to write about. DSP became accessible to mid-sized brands rather than only enterprise accounts. Amazon Marketing Cloud made it possible to reconcile DSP and sponsored ads against each other instead of stacking two dashboards and hoping. Streaming TV inventory turned into a real line item rather than a novelty. The strategy question worth answering now is not "which format should I try." It is "how do I split budget across the funnel, and how do I know any of it moved sales that would not have happened anyway."

So this page keeps the parts of the 2020 playbook that still hold, drops the parts that do not, and spends most of its words on the question the 2020 pages never asked: whether the spend is incremental.

The 2020 Playbook, Format by Format: What Still Works

Here is the honest audit of the 2020 playbook, format by format.

• Sponsored Products — still the workhorse. Keyword or product targeting, cost-per-click, one ASIN per ad. The mechanics did not change; what changed is that it is now table stakes rather than an advantage.
• Sponsored Brands — a headline, a logo and up to three products above or beside search results, still billed per click. In 2020 the differentiator was having a registered brand. Now the video version is the one worth testing first.
• Sponsored Display — retargeting and competitor-detail-page placements, on Amazon and off, priced per click or per thousand impressions. Same job as 2020; it just gets reported alongside DSP now instead of separately.
• Amazon DSP — programmatic display, video and streaming TV bought on a CPM basis across Amazon.com, Fire TV, IMDb, Twitch and third-party sites. In 2020 this was an enterprise conversation. Now it is a mid-market one, and the barrier is measurement literacy, not access.
• Stores — a free multi-page brand storefront. Still free, still the most underused page in most accounts, and still the cheapest place to send DSP traffic that is not ready to buy yet.
• Posts — the free shoppable feed the 2020 guides recommended. It is not a channel you can build a 2026 plan on.
• Vine — Amazon's own review programme for new ASINs. Still a legitimate launch tool, still worth using on a new product, and still not a strategy.

Read that list and the pattern is obvious: every item on it is inventory. None of them is a plan. A plan is a decision about which shoppers you are paying to reach, what you are willing to pay to reach them, and how you will know whether it worked.

The Part That Hasn't Changed: Where the Budget Goes

Sponsored Products and Sponsored Brands are bottom-funnel. They catch demand that already exists — someone searched, someone browsed a similar product, someone is close to buying. DSP, particularly display and video, is top and mid funnel. It builds the audience that eventually searches. Confusing the two is the most common strategy error we see: brands park 100% of budget in Sponsored Products because it is the easiest thing to measure, then wonder why growth stalls once branded search demand plateaus.

How much should go to DSP versus sponsored ads is a genuinely bad question to ask in the abstract. It depends on catalog size, review count, and how much of your revenue is already branded search versus category search. A brand with under 30 reviews on its hero SKU has a review problem to solve before it has a media-mix problem. A brand with a mature catalog and flat branded search growth is underinvesting at the top of the funnel, full stop.

The single number that tells you which side of that line you are on is new-to-brand share. If it is healthy and rising, your media is finding people. If it is low and falling while ROAS holds steady, you are paying to re-serve ads to shoppers who were going to buy anyway — which flatters the dashboard and does nothing for growth.

A Worked Example: What a Mature DSP Programme Actually Produces

Numbers help more than percentages, so here is a real one. Across 30 advertisers we manage, in July 2026, the portfolio delivered 6.04x return on ad spend — measured across the whole book, not cherry-picked from the best line item. That book ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost-per-click.

The $0.41 CPC that circulates in this category is a real number, but it is online-video only, not the whole book. Compare it against a blended figure and every real campaign looks expensive next to something that was never comparable. At a $4.00 CPM, a million impressions costs $4,000 — which is why the CPM matters far less than what those impressions go on to do.

Blended cost per acquisition across that same book was $5.49 across 57,137 attributed purchases, and 20.1% of those purchases — one in five — came from a shopper new to the brand. That last figure is the one that tells you whether DSP is doing top-of-funnel work or taxing an audience you already had. None of this is a promise about your account. It is what one large, mixed portfolio produced in a single month, and it is most useful as a sanity check on whatever benchmark someone quotes you.

The Measurement Mistake Almost Everyone Makes

Last-click attribution credits whichever ad the shopper clicked last. A DSP display ad that nudged someone toward a purchase gets zero credit if that shopper later clicked a Sponsored Products ad on the way to checkout. Add the DSP report and the sponsored ads report together and you will double-count the same sale. At scale that is not a rounding error — it is the difference between a campaign that looks like it is working and one that is proven to work.

We have made this mistake ourselves, early on, running the same last-click dashboards everyone runs before Amazon Marketing Cloud made it possible to reconcile the two. The honest fix is a holdout or matched-control test: show the ad to part of your audience, withhold it from a comparable group, and measure the actual lift. Last-click can tell you a sale happened. It cannot tell you the ad caused it. Those are different questions, and only one of them is worth building a strategy on.

A 2026 Checklist, and What to Check When the Numbers Look Bad

If you want to replace a 2020 plan with something current, work through this in order. It is roughly the sequence we use.

• Fix the listing before the bid. If clicks are healthy and sales are not, no bid change fixes it — images, price, review count, A+ content.
• Pull the search term report and negate. Broad and automatic targeting spend against searches that have little to do with the product. This is usually the fastest single win in an account.
• Set a break-even ACoS from your margin, not from a benchmark. What is left per order after product cost, fees and shipping, divided by price, is your ceiling. Anything else is guesswork.
• Decide whether you have a reach problem or a conversion problem. Flat branded search and a mature catalog means reach. Thin reviews mean conversion.
• Only then add DSP, and only with a holdout or an Amazon Marketing Cloud reconciliation in place before the first impression runs.
• Track new-to-brand share monthly, not ROAS alone.

When the numbers look bad, check the boring things first. Amazon has adjusted attribution windows more than once, and a shortened window makes identical performance look like a decline. A CPC that looks high against a benchmark you found online may be being compared to a video-only figure. And if your DSP runs through an agency-owned seat rather than one you control, ask who owns the audience data and reporting access before you troubleshoot anything else — you may be diagnosing a visibility problem, not a performance one.

Where a Managed DSP Partner Fits, and Where It Doesn't

If your catalog is small, your reviews are thin, or you have not yet got Sponsored Products to a stable target ACoS, a DSP strategy is premature. Fix the bottom of the funnel first, and you can do that in-house without a DSP seat at all.

DSP earns its place once you have a catalog and a review base that can carry top-of-funnel spend, and you actually need to prove whether that spend is incremental rather than just visible. reMKTR runs Amazon DSP as a managed service on our own advertiser accounts rather than reselling access, and we reconcile DSP against sponsored ads in Amazon Marketing Cloud so the same sale does not get counted twice. We measure with holdouts and matched controls, not last-click. We are part of the Full Circle group (100+ brands managed, $500M+ in Amazon revenue managed). None of that makes DSP the right next move for every brand reading this — but if the question you are trying to answer is whether your display spend is doing anything, that is the question our reporting was built for.