Acorn alternatives — four options, and the handover nobody plans for

Two words that send people to the wrong shortlist

Before the options, clear two things that waste a great deal of evaluation time.

"Acorn" is four companies. The one relevant here is Acorn-i at acorn-i.com — a London ecommerce agency describing itself on its own site as "A Technology Led Ecommerce Agency" and "Experts in Amazon and ecommerce", running an analytics application called Ignite and publishing a 94% client-retention claim beyond twelve months. It is not Acorn by New Engen, not Acorns the investing app, not Acorn TV. If your shortlist has picked up any of those, start again.

"DSP" is two things. Amazon DSP is the Demand-Side Platform — the programmatic system for buying display, video and audio inventory on and off Amazon with Amazon shopping signal behind it. Amazon's Delivery Service Partner programme is the parcel-delivery franchise. They share three letters and nothing else.

One more thing worth knowing before you compare anything. Acorn-i's own site does not name Amazon DSP. That is not a criticism — plenty of capable agencies buy DSP without featuring it on the homepage — but it does mean the first question in your replacement search is whether DSP was in your scope at all, or whether what you actually had was sponsored ads plus content. Those are different replacements. Get the last twelve months of your own invoices out and check which line items existed before you shop for a substitute for them.

Category one — another full-service commerce group

The like-for-like move: replace one agency with another that also does content, insight and media across several retailers and several markets.

Who this genuinely suits. Brands whose real job is coordination. If you sell in five countries across Amazon, a DTC site and one or two other marketplaces, and the thing that breaks is that nobody owns the whole picture, then the answer is a group who can hold it — and the media is one workstream inside that, not the point of the exercise. It also suits brands with a thin internal team, where the agency is effectively the ecommerce department.

What to watch. Full-service groups vary enormously in how deep the DSP bench goes, because display is a specialist craft and a generalist agency may run it competently rather than expertly. Ask to meet the person who will be in the console, not the person who runs the account. Ask what their median advertiser returned last month and how wide the spread around it was — not the portfolio average, which a handful of deep retargeting pools can carry on their own.

The consolidation question. This part of the market has been changing hands steadily for four years: platforms folding into suites, agencies folding into holding companies, software vendors taken private. That is not a warning about any particular firm and we are not going to imply one. It is a reason to put change of control into the contract conversation: if the company is acquired, what happens to your rate, your team, and your notice period? Every buyer should ask it, of every candidate, including us.

Category two — licence the platform and run it yourself

Bring it in-house: buy a retail media platform, hire or redeploy a trader, and own the console.

This is the option most shortlists include and fewest shortlists price properly, so here is the arithmetic that decides it.

Skai publishes a full rate card — read on skai.io on 20 August 2026, in US dollars, annual, banded by annual media spend: $114,000 up to $4M, $276,000 up to $10M, $504,000 up to $20M, $756,000 up to $35M, custom above. Now set that against Amazon's own published floor: Amazon's DSP page states that its managed-service option "typically requires a minimum spend of $50,000", varying by country.

A brand sitting at roughly that managed-service floor is running in the region of $600,000 a year of media. A $114,000 licence against $600,000 of media is 19% of the entire budget — before you have paid anybody to operate it. The same licence against $4M of media is 2.85%. That single comparison is why category two is not really a live option for most brands: a published platform licence only starts making sense somewhere in the low millions of annual media, and below that the arithmetic closes the door regardless of how appealing in-house control sounds.

Add the second cost nobody puts in the model: a competent trader, fully loaded, plus the time it takes them to become competent in a specific console. And note one more thing from that same Skai page, as a category fact rather than a swipe — incrementality testing appears from Enterprise Premier upward, so the ability to prove advertising caused the sale can itself be a tier rather than a default.

Who this genuinely suits: brands above roughly $4M of annual media with an existing programmatic team, or a strong strategic reason to keep the buying knowledge inside the building. For those brands it is a good option and often the right one.

Category three — a specialist Amazon DSP buyer

Narrower than a full-service group and deeper: a partner whose whole business is Amazon retail media, holding their own DSP seats, with people who buy display every day.

Who this genuinely suits. Brands where Amazon is the main event rather than one of five channels, where the DTC site and the content are already handled, and where the open question is whether display is buying incremental sales or repurchasing shoppers you already had. Specialists tend to be materially better at that question than generalists, because they have run the test more times.

What to ask that generalists rarely get asked. What did your median advertiser return last month, and how wide was the spread around it? What is your typical CPM and blended cost per click across your book, and over what period — and will you scope it honestly rather than quoting the best line item? Do you run audio and video as well as display, or display only? How do you reconcile DSP against sponsored ads so the two stop taking credit for the same order?

Add one more, because it comes out of our own numbers and almost nobody asks it. Across 27 advertisers over the 31 days from 15 July to 14 August 2026, off-Amazon exchange inventory returned 6.05x against 4.90x on Amazon's own shopping surfaces — and found far fewer new customers, 15.9% new-to-brand against 42.0%. Amazon-owned shopping placements took $439,875 of that spend and third-party exchange inventory $259,516. The two are doing different jobs: the exchange converts people already in the market, the owned surfaces recruit. So ask a specialist what target they set for each, and whether they set one number for the whole account. A single blended return target quietly starves acquisition, because the half that recruits will always look like the worse half.

That last question is where most of the money is. Sponsored ads and DSP can both claim the same purchase, and an account looking at two dashboards will happily believe both. The reconciliation happens in Amazon Marketing Cloud, where event-level data from both sits together, and a partner who cannot describe how they do it is telling you something.

The honest limitation of this category. A specialist will not fix your product photography, your listings, your international expansion or your DTC funnel. If those are the constraint, buying deeper media expertise is buying a better engine for a car with a flat tyre.

There is also a concentration question worth asking specialists that generalists are rarely exposed to. A firm whose entire business is one retailer's advertising products is, by construction, exposed to that retailer changing them — and Amazon changes them regularly. That is not a reason to avoid specialists; depth is exactly what you are buying. It is a reason to ask how the firm handled the last significant change to Amazon's DSP or measurement products, what they did in the following month, and how they told their clients. The answer separates firms who read release notes from firms who wait for their reporting to break.

Category four — Amazon DSP delivered as a managed product

The fourth shape is neither an agency retainer nor a licence: the DSP run for you, daily, as a product with a defined scope and a defined approval model. That is what Dr. DSP is, so treat this section as the interested party's account of itself and score it on the same questions as the rest.

Fable 5 reads the account and works it every day, with operators from Full Circle supervising — a full-service Amazon management company with more than $500M in managed revenue across 100+ brands. You set 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. Orbit, the full software suite, is included at no extra cost — and Orbit is not a DSP, so do not read it as a substitute for the platform.

The measurement position, stated carefully: last-click attribution cannot prove incrementality, because it can only see conversions it already claimed. Holdouts and matched controls can. We design the test before the budget moves and reconcile DSP against sponsored ads in Amazon Marketing Cloud. Several serious competitors now advertise incrementality testing too, and it would be dishonest to imply otherwise — the question that actually separates vendors is who designs the test, who computes it, and whether it is named in the statement of work. Ask us that as hard as you ask anyone.

The book behind it: across 30 advertisers in July 2026, a live Amazon DSP API pull showed 6.04x return on ad spend, 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost per click, with a $5.49 cost per acquisition across 57,137 attributed purchases and 20.1% new to brand. One month, thirty advertisers, scoped exactly as measured.

Where we lose. We publish no price — Skai does, and on transparency they are ahead of us. We do not produce your creative assets or run your DTC site. And if the reason you are leaving is that nobody coordinated your five channels, a media product is not the coordinator you need.

The handover, in the order it has to happen

Switching partners is not a decision, it is a sequence, and the sequence matters more than the choice. Nine things move, and only one of them is irreversible.

• The advertiser seat. Establish whose it is before anything else. If it belongs to the outgoing partner, you are not transferring an account, you are building a new one. This is the irreversible one.
• Amazon Marketing Cloud instance access — who holds it, what queries exist, and whether the historical data stays reachable.
• Audience definitions — the logic, not just the names, so they can be rebuilt identically.
• Creative assets and their approvals.
• Historical performance, exported at line-item level, before you give notice rather than after.
• Conversion and pixel configuration, documented.
• Supply and inventory exclusions — the blocklists somebody spent a year building.
• Naming conventions, so next quarter's reporting can still see this quarter.
• The measurement baseline — agree with the incoming partner what "before" looks like, in writing, on the day they start.

Now the parallel-run arithmetic, which is the part nobody costs. Most handovers involve sixty days of overlap, and during that overlap you are paying two parties against one media budget. If both fees come out of the media rather than being invoiced beside it, the duplication is real spend. At a 5.70% implied rate — Skai's published Standard tier against $2M of annual media — sixty days of overlap on a $100,000-a-month budget duplicates about $11,400 of fee. The useful way to feel that is not in dollars. At the $4.00 CPM we measured across 30 advertisers in July 2026, $11,400 is roughly 2.85 million impressions you paid for and did not receive.

That is not an argument against overlapping — a clean handover is usually worth it. It is an argument for asking one question of both partners in writing: is my media budget inclusive or exclusive of your fee during the transition? Get it exclusive if you can, and the overlap costs you a fee instead of costing you reach.

Before you replace anything, find out which channel is leaking

A fair share of brands shortlisting a new display partner do not have a display problem. Two figures from our own data make the difference legible, and they fail in opposite directions.

On the sponsored-ads side, read across the 47 Amazon accounts our group manages, 48.5% of all search spend — $4,962,963 of $10,243,379 since 1 May 2026 — went to search terms that produced no orders at all, and 83% of every term that took a click returned nothing. Nobody chose those terms. They arrived through matching, and they kept arriving.

On the display side the waste has the opposite shape. In a 31-day pull of our own DSP book — 27 advertisers, 15 July to 14 August 2026, a different window and a different advertiser set from the July figures elsewhere on this page, so the two are not to be read against each other — one in five spending line items produced zero attributed purchases, and between them they burned 0.7% of the budget: 259 of 1,325 line items, about $4,927. At the other end, 18.3% of line items — 242 of them — took 80% of the spend.

Read together, those two give you a diagnosis before you take a single call. On DSP you choose every line item, so the failures are few, small and nameable, and the money sits in a couple of hundred places you could list on one page. On search the failures are automatic and enormous. If your waste is search-shaped, replacing your display partner will not touch it and Dr. PPC is the faster fix. If your budget is concentrated in a handful of line items nobody has re-read this quarter, that is a display problem, and the four categories above are the choice in front of you. The two channels need entirely different waste controls, which is also why one partner rarely fixes both by accident.

Which of the four you should actually take

Short version, by the thing that is actually broken.

• Nobody owns the whole picture across channels and markets — take another full-service group. Media depth is a secondary criterion; coordination is the job.
• You have a programmatic team and more than roughly $4M of annual media — licence a platform. Below that the arithmetic above closes the option, whatever the demo felt like.
• Amazon is the main event and you want people on it — take a specialist. Ask for seat count, blended CPM and CPC over a stated period, and how they reconcile against sponsored ads.
• You want the work done daily and the incrementality question answered rather than asserted — that is the managed-product shape, and it is what we built.

Then, whichever you take, read the contract before the feature grid. Quartile publishes theirs and it is the clearest worked example in the category: a one-year initial term auto-renewing for successive one-year periods, sixty days' written notice of non-renewal, fees due in advance and non-refundable, a sixty-day evaluation window at the start during which you can terminate from inside the product, and month-to-month available only as a right they "may, in [their] sole discretion, grant certain Clients". That is a normal enterprise contract and publishing it is to their credit. It is also the reason your notice date belongs in a diary the week you sign.

Two places to go next if this did not settle it. If what you actually need is the fee architecture rather than the shortlist, our page on Acorn-i cost works through the four shapes and what one undisclosed point costs. If you want the head-to-heads scored against named competitors, Acorn vs the field has the seven-question scorecard. And if it turns out display was never the constraint and sponsored ads are where the waste is, Dr. PPC publishes its price and starts there.

Dr. DSP and Dr. PPC are products of Full Circle, which has managed more than $500M in revenue across 100+ brands.