Skai vs Pacvue — two good platforms built from opposite ends of the problem

The short version, then the argument

If your organisation's centre of gravity is a media team that already buys search and social and now needs retail media inside the same workflow, Skai is the more natural fit. If your centre of gravity is a commerce team that lives in retailer operations — availability, buy box, content, catalogue — and wants advertising attached to that, Pacvue is. Both will demo well. The one your people actually adopt is the one that matches how they already work.

That is a genuinely unsatisfying answer if you were hoping for a winner, so the rest of this page makes it concrete: where each came from, the one dimension on which they differ sharply, a test you can run internally in an afternoon, and the problem neither of them solves for you.

One clarification because this page borders on demand-side platform territory. DSP means demand-side platform — software for buying programmatic display, video and audio. Amazon's Delivery Service Partner programme, which franchises parcel delivery, shares the letters and nothing else. This comparison is written by Dr. DSP, an Amazon demand-side platform product from Full Circle, a full-service Amazon management company with more than $500M in managed revenue across 100+ brands. We are not a neutral party and we would rather say so at the top than pretend otherwise.

Where each came from, because it is still visible in the product

Skai began life as Kenshoo, founded in 2006 in Tel Aviv, built around paid search and performance marketing for large advertisers. It renamed to Skai in June 2021, following its acquisition of Signals Analytics, as it pushed further into commerce and data intelligence. That heritage shows: breadth across publishers, an emphasis on cross-channel management, and the instincts of a bid-management platform applied to retail media.

Pacvue came from the other direction — a Seattle company built for marketplace advertising and commerce operations from the start, acquired by Assembly, a portfolio company of Advent International and PSG Equity, in October 2021 in a deal whose terms were not disclosed. Its platform pages describe a product organised around media, commerce operations, revenue recovery, market insights, a data-as-a-service layer and an AI agent, with claimed support across 100+ retailers in 30+ countries.

Neither origin story is better. They explain what each product is unusually good at and what each treats as an adjacency. A platform built to manage bids across publishers thinks in campaigns and channels; a platform built for marketplace operations thinks in ASINs, availability and share of shelf. Both eventually offer most of the same features, and the difference persists in the parts you cannot see on a feature grid — defaults, hierarchy, and what the first screen shows you.

The one dimension where they genuinely diverge

On price transparency the two sit at opposite ends, and this is worth stating precisely because it is the only hard, checkable difference a buyer can verify without a sales call.

Skai publishes a list. Read on 20 August 2026, its pricing page shows annual tiers banded by yearly ad spend through the platform: Standard at $114,000 for advertisers up to $4M of annual spend, Advanced at $276,000 up to $10M, Enterprise at $504,000 up to $20M, Enterprise Premier at $756,000 up to $35M, and a tailored band above that. The page also mentions flexibility to review the commitment after the first three months, and lists incrementality testing at the Enterprise Premier tier.

Pacvue publishes no figure. Its pricing URL returns a 404 and the site routes to a demo booking, which is the norm in this category rather than an outlier. Both approaches are legitimate. What matters for you is the asymmetry it creates in a bake-off: you will be comparing one known quantity against one unknown, and unless you force both into the same unit — total annual cost at your real spend, with scope held constant — the comparison will quietly favour whichever one you understood better.

Do the conversion before either demo. Write down your true annual spend, the retailers you need, the number of seats, and whether you need programmatic display as well as sponsored ads. Then ask both vendors to price that identical brief.

A test you can run internally before you talk to either

Adoption kills more platform purchases than capability does. Run this in an afternoon and it will predict the outcome better than a feature matrix.

• Ask who will log in daily. Not who will own the contract — who opens the tool on a Tuesday morning. If you cannot name that person, no platform on this shortlist will succeed and you are actually shopping for a service.
• Ask what they open first today. A media planner opening a campaign dashboard and a commerce manager opening an availability report want different first screens.
• Count the retailers that matter. If Amazon is more than the large majority of your revenue, breadth across 100+ retailers is a feature you will pay for and not use.
• Count the channels outside retail media. Substantial paid search and social budgets that need managing alongside retail media pull hard toward a cross-channel platform.
• Establish who does measurement. If nobody in-house can write a clean-room query, a platform that automates that work is worth more than one that exposes it.
• Check the calendar honestly. How many hours a week will this genuinely get? Under about five, buy a service instead. Software is cheap and unattended software is expensive.

The problem neither of them solves for you

Both platforms will improve what you can see. Neither can tell you, on its own, whether the improvement was real — and that is not a criticism of either product, it is a property of the whole category, ours included unless we design around it.

Last-click attribution cannot prove incrementality. Any system optimising toward reported conversions will drift toward the buys that report best, and in display that means retargeting shoppers who were already returning. The reported return climbs. The business does not move. Everyone congratulates the platform.

The correction is not a better dashboard, it is a designed test: withhold display from a matched set of regions or ASINs, run a defined window with a stated failure condition, and reconcile the result in Amazon Marketing Cloud, where DSP impressions and sponsored-ads events land in the same clean room and stop each claiming the same order. Pacvue's own platform copy makes a related promise — one performance truth tied to business outcomes rather than last click — and Skai lists incrementality testing in its top published tier. Both are real signals that the vendors take this seriously. Press on the specifics anyway: which holdout method, what minimum spend and duration for a readable result, who writes the test, and can you see an anonymised output. Amazon offers AMC free to eligible advertisers, so the limiting factor is analyst time and willingness rather than licence cost.

When neither is the right purchase, and what we do instead

There is a third answer that belongs on this page for honesty's sake: sometimes the correct decision is not to buy a platform at all. If Amazon is effectively the whole business, if no one internally has advertising as their actual job title, and if the open question is whether display is producing sales that would not otherwise have happened, then you are shopping for an outcome and a console will not deliver one.

Dr. DSP is Amazon display run as a managed product. Every proposal carries the evidence behind it, how it will be measured, and the condition that automatically reverses it. You pick the autonomy level — every change waiting on your approval, routine work automatic with the bigger moves queued, or fully autonomous inside agreed guardrails — and you can move it at any time. Inventory risk, pricing, new products, new creative and the decision to stop spending on display always come to a human, at every level. There is no published price: a demo, the first 30 days free, priced on the call against real media budget and scope, month to month, with Orbit included at no extra cost.

For a yardstick rather than a promise: across 30 advertisers in July 2026 — not the whole book — a direct API pull showed 6.04x return on ad spend, 78.4 million impressions at a $4.00 CPM, a blended $1.42 cost per click, $5.49 cost per acquisition across 57,137 attributed purchases and 20.1% new-to-brand. Amazon-attributed, one month, guaranteed to nobody.

Two redirects that may fit better than we do. If sponsored ads rather than display are where the money is going astray, Dr. PPC is the right product and it publishes its price and its cap. If your reporting keeps getting distorted by stockouts and fee leakage rather than by bidding, fix that first with Dr. Stock — no platform comparison is worth running while inventory is the real variable.