Quartile review — read the published numbers, then read the contract

What Quartile is, in its own words

Read on quartile.com on 20 August 2026, the company describes itself as "The World's Largest Retail Media Optimization Platform", combining proprietary algorithms with strategic expertise to drive full-funnel growth for ecommerce brands.

The channel list is broad and, unusually for this category, it names Amazon DSP explicitly alongside Amazon sponsored ads, Walmart, Instacart, Criteo — which is how it reaches Target and Best Buy inventory — plus Google, Meta and Microsoft ads. That deserves noting because several serious competitors in retail media do not name Amazon DSP on their own product pages at all, which leaves buyers guessing about scope. Quartile does not leave you guessing.

A disambiguation for anyone who arrived from a search that could go either way: Amazon DSP is the Demand-Side Platform, Amazon's programmatic system for buying display, video and audio inventory with Amazon shopping signal behind the targeting. It is not the Delivery Service Partner parcel franchise.

So the product is: a bidding and optimisation layer applied consistently across many retail networks, with an account team attached, sold to brands and agencies with large catalogues. It is not an agency, and it is not a self-serve tool you buy with a credit card. It sits in the middle, and understanding that is most of what a fair review needs to establish.

This review deliberately contains no star ratings and no review counts. The major review sites block automated reading, and search snippets of those pages have contradicted the live pages more than once in our own checking. Printing a number we could not open ourselves would be exactly the error this page is trying to help you avoid.

Four published numbers, and the arithmetic nobody runs on them

Quartile publishes four scale figures on its homepage: $2B+ in annual retail ad spend managed, $25B+ in annual managed sales, 5,300+ customers globally, across 32+ countries.

Those are impressive on their own and they are almost always quoted as-is. Divide them by each other, though, and they answer the question a prospective buyer actually has — is this product built for a business like mine? — which the homepage never addresses directly.

• $2B of managed ad spend across 5,300+ customers averages about $377,000 of annual media per customer — in the region of $31,400 a month.
• $25B of managed sales across the same base averages about $4.7M of annual sales per customer.

Both are averages across a book that certainly contains very large accounts and much smaller ones, so nobody should read them as a threshold or a minimum. What they legitimately give you is a centre of gravity, and it is genuinely useful in two directions.

If you run $8,000 a month of retail media, you sit well below that centre. That is not a reason to rule Quartile out — plenty of platforms serve customers below their average happily — but it is a reason to ask directly what service level and what account coverage that spend buys, and to get the answer before the contract rather than in month four.

If you run $300,000 a month, you are an order of magnitude above the average customer, and you have considerably more negotiating room than the sales conversation will imply. Very few buyers work that out, because very few buyers divide the two numbers on the homepage.

For a sense of scale on the other side: Amazon's own DSP page states that its managed-service option "typically requires a minimum spend of $50,000", varying by country. Quartile's average customer sits below that, which tells you the book is weighted towards sponsored ads and multi-retailer work rather than pure display.

The 12.5x figure, read carefully — including against ours

The second piece of arithmetic those four numbers permit is the one most likely to be misused, so here it is with the cautions attached before the figure.

$25B of managed sales against $2B of managed spend implies a blended 12.5x across Quartile's whole published book — equivalently, an 8% advertising cost of sales. Quartile also publishes a claim of a "41% average increase in ROAS when using Quartile", which is their own figure, self-reported, and we have no way to verify it. We are reporting that they publish it, not endorsing it.

Now the cautions, and they matter more than the number.

It is not a like-for-like comparison with anything. Managed sales across sponsored ads, DSP, DTC and several retail networks is a much broader denominator than DSP-attributed sales alone. Blended figures also flatter any book weighted towards sponsored ads on branded search, where returns are structurally high and incrementality is structurally low. None of that is a criticism of Quartile — it is a description of what a company-wide blended multiple is, wherever it appears.

Including on ours. Our own live Amazon DSP API pull across 30 advertisers in July 2026 showed 6.04x return on ad spend — measured across the whole slice rather than cherry-picked from the best line item, and covering display specifically. That is lower than 12.5x. We are saying so rather than leaving you to notice. The honest position is that the two numbers are not measuring the same thing and neither should be used to rank the two companies: theirs blends channels including search, ours is DSP across one month for thirty advertisers.

The point of the exercise is not to win a comparison. It is that a blended multiple — anybody's — cannot tell you whether the advertising caused the sales. That is the next section, and it is the one that should decide your shortlist.

The question a bidding platform structurally cannot answer for you

Bid optimisation makes each dollar buy better placements. It is real, valuable work and Quartile does it at very large scale across more networks than most competitors touch.

What it does not do — because no bidding layer does — is tell you whether a given order would have happened without the advertisement. Last-click attribution cannot prove incrementality: it observes a click or a view followed by a purchase and assigns credit, with no representation of the purchase that would have occurred anyway. Optimising against that signal makes you very efficient at buying credit for sales you already had.

The reason this bites hardest on display rather than search is a matter of who is being reached. Across 30 advertisers in July 2026, 20.1% of our attributed purchases were new to the brand — which means roughly four in five came from shoppers who already knew it. That larger share is precisely where the counterfactual question is hardest and where the money is most likely to be re-authorised on faith.

What answers it is a holdout or a matched control, designed before the budget moves: withhold the audience from a randomised slice, or match a comparable untreated group, and measure the difference rather than the total. For Amazon that requires ad and retail events together, which is Amazon Marketing Cloud — also the only sane place to reconcile Amazon DSP against sponsored ads so they stop claiming the same orders.

Be scrupulous about who else offers this, because the temptation to overclaim here is exactly where our own past errors have come from. Criteo advertises incrementality testing on its own advertiser pages. Tinuiti runs a published incrementality practice. Skai lists incrementality testing on its own published rate card from Enterprise Premier upward. This is not a capability anybody invented recently and it would be dishonest to imply a competitor lacks it.

The question that actually separates vendors is who designs the test, who computes it, and whether it is named in the statement of work — because a test designed and computed by the party being measured is a different instrument from one that is not. Ask Quartile. Ask us harder.

The contract is the product — and Quartile publishes theirs

Here is the most decision-relevant thing about Quartile, and it appears in no ranking review of the company: they publish their full terms of service. Most of their peers do not. Read them before you read another feature comparison.

Verified on quartile.com/legal/terms-and-conditions on 20 August 2026:

• Initial term of one year from the effective date, which "shall automatically renew on each anniversary of the Effective Date for successive one-year periods".
• Sixty days' written notice of non-renewal, by either party, before the end of the then-current term.
• Fees due in advance of services rendered, unless the order form says otherwise.
• No refund "in connection with any termination of this Agreement".
• A defined sixty-day Evaluation Period from the effective date, during which the client can terminate using a designated button.
• Terminating inside the term for any reason other than Quartile's uncured material breach means paying all fees that would have been due for the remainder of the term.
• Month-to-month exists — but as a right Quartile "may, in its sole discretion, grant certain Clients", with thirty days' notice.
• Pricing tiers are set by "Client's aggregate monthly advertising spend", "the number of advertising platforms that Quartile manages", and/or "the level of Service agreed upon".

None of that is improper. It is an ordinary enterprise software contract, and publishing it openly is genuinely to Quartile's credit — you can read it before the call, which is more than most of this category permits.

What it means in practice, as arithmetic. Of a 365-day year, 60 days are freely reversible — 16.4% of the term — and the remaining 305 days are committed unless there is a material breach. Miss the notice date by a single day and the whole 365 repeats. That last clause is the one that catches people, and the fix costs nothing: put the notice date in a calendar the week you sign, with a reminder ninety days ahead of it so you have time to run a decision rather than a panic.

The other clause worth negotiating rather than accepting is the month-to-month one. It exists in the document. Whether you get it is a discretionary grant, which means it is a thing to ask for at signature, when you have leverage, rather than at renewal, when you do not.

Reviews you cannot read, and a thirty-day evaluation you can run

Two structural problems make written reviews of any platform in this category less useful than they look, and both have fixes.

Reviews predate the products. Retail media tooling has been rebuilt repeatedly in three years — platforms renamed, tools absorbed into suites, AI layers added, ownership changed hands. A review from 2024 frequently describes an interface and a scope that no longer exists. Date every review before you weight it.

The review sites are unreadable to automated checking. We could not open them, so we are not going to quote them. What we will say is that any figure lifted from a search snippet is unreliable — snippets have contradicted the live pages more than once in our own research.

What to do instead, and it is better evidence anyway:

• Two reference calls with customers near your own spend level — one of them at least two years in. Ask what has changed since they signed, and how long a genuine escalation took.
• Ask for the last four weeks of change logs on a comparable account. Cadence is what reviews capture worst and what determines results most.
• Run a thirty-day structured evaluation rather than a demo. Quartile's own terms give you sixty days of reversible time, so use the first thirty of them deliberately: name three products, agree the baseline in writing on day one, agree what would count as success on day one, and change nothing else in the account for the duration.
• Ask what happens at renewal — specifically what notice you get before a fee change. Get that in writing, from any vendor including us.

The complexity criticism that follows platforms of this kind around is worth taking seriously in a specific form rather than a general one: a system that spans eight ad networks has an irreducible learning curve, and the honest question is not whether it is complex but who inside your business will own it, and how many hours a week they have. If the answer is nobody with fewer than five, that is a fit problem no feature will solve.

Where Dr. DSP differs, and who should ignore us

Who should ignore us, first. If you sell meaningfully across Walmart, Instacart, Target and Best Buy as well as Amazon, and you want one optimisation layer applied consistently across all of them, Quartile is built for that and we are not. We are Amazon-first by design. A multi-retailer brand choosing us would be choosing a narrower product on purpose, and there needs to be a reason.

Where we differ. Dr. DSP is Amazon DSP run as a managed product rather than a platform you drive. Fable 5 reads the account and works it daily; operators from Full Circle supervise — a full-service Amazon management company with more than $500M in managed revenue across 100+ brands. You choose the autonomy level: every change waiting on your approval, routine changes automatic with the larger ones queued, or fully autonomous inside agreed guardrails. Every proposed change carries the evidence behind it, a measurement plan and a rollback trigger before it runs.

The measurement sits at the front rather than the end. The holdout or matched control is designed before the budget moves, and DSP is reconciled against sponsored ads in Amazon Marketing Cloud so the two stop taking credit for each other. Orbit, the full software suite, is included at no extra cost — and Orbit is not a DSP, so it is not a substitute for the platform.

Where we lose. Quartile publishes its terms of service and its scale figures. We publish no price for Dr. DSP at all, and Skai — with four exact annual tiers on its own page — is ahead of both of us on that. Quartile covers retail networks we do not touch. And their published blended multiple is higher than the 6.04x we measured across 30 advertisers in July 2026, even though the two figures are not measuring the same thing.

Two places to go next. If the fee architecture rather than the product is your open question, our breakdown of Quartile pricing works through the tier structure their terms describe. And if your spend is mostly sponsored ads today with display still a question, Dr. PPC publishes its price — $300 a month plus 3% of ad spend, capped, month-to-month, first 30 days free — and is the cheaper place to begin.