Teikametrics review: judge it on the marketplaces you actually sell on

What you are reviewing now, and why the name matters

Teikametrics presents its platform as ARI — Artificial Retail Intelligence — described on its own site as an "AI Marketplace Optimization Platform" and "The Marketplace Intelligence Platform That Scales With You". The older Flywheel name survives mainly in legacy sign-in URLs.

That matters for a practical reason rather than a cosmetic one. When a platform is renamed, the searchable record splits: the old name carries the reviews and the new name carries the product. You end up reading detailed accounts of an interface, a bidding model and a support experience that may or may not describe what you would be buying. Nobody is at fault — the writers were accurate when they wrote — but the mismatch is real and it is your problem to manage.

The habit worth forming, and it applies to every vendor including us, is to open the vendor's current product page in one tab and the review in another, and check that the review's feature claims still exist. Two minutes of that filters more noise than any amount of star-rating arithmetic. It is also the only workable approach right now, because the major review platforms block automated access and we will not quote a rating we cannot verify against the live listing.

What has not changed is the shape of the product: goal-driven automated bidding across marketplaces, with reporting and intelligence layered over it, sold as software you operate rather than a service that operates it for you.

The strongest reason to buy it: Walmart and TikTok Shop are not afterthoughts

Plenty of Amazon tools list Walmart on a feature grid. Far fewer treat it as a marketplace with its own auction dynamics, its own reporting lag and its own optimisation logic. Teikametrics has been building multi-marketplace for years, and its current site names Amazon, Walmart and TikTok Shop as the channels it operates across.

If you genuinely sell across those, that is the reason to shortlist it, and it is a real one. Running three marketplaces from three consoles with three separate mental models is a tax that compounds: your team learns three sets of quirks, your reporting never reconciles, and the budget conversation happens three times with no shared denominator. A platform that flattens that is worth paying for.

Here is the test that should decide it, and we have not seen anyone write it down. Take the share of your revenue that sits outside Amazon. If it is under roughly 10%, a multi-marketplace platform's central advantage is close to unusable for you — you would be buying an Amazon tool at a multi-marketplace price, and the honest comparison set is Amazon-only tools. Between 10% and 30%, it becomes a real question about where your growth is going rather than where your revenue is today. Above 30%, the unified view is probably worth more than any bidding-algorithm difference between vendors, and you should weight it accordingly.

That single ratio settles more of this decision than a feature comparison will, and it takes five minutes to calculate from figures you already have.

What automated goal-driven bidding decides, and what it never touches

Every automated bidding platform, this one included, is very good at one class of problem and structurally blind to another. Being clear about the boundary is more useful than praising or criticising the algorithm.

What it does well: move bids toward a target continuously, at a cadence and consistency no person sustains; hold a goal across thousands of keywords; and stop the slow drift that turns a well-built account into an expensive one over six months.

What it does not do is decide whether the structure underneath is right. Consider the scale of what sits below the bidding layer. Across the 47 brands we manage, Amazon delivered clicks on 3,086,624 distinct search terms from 1 May 2026 onward, and 83% of them took money and returned nothing.

An automated bidder will reduce what it spends on those terms as the evidence accumulates, which is exactly what it should do. What it cannot do is notice that two of your campaigns are competing for the same phrase, that a converting term is throttled by a budget you set eleven months ago, or that a product should not be advertised at all this month because the margin has moved. Those are judgement calls about account architecture, and they are made by people or not at all.

This is not a criticism of Teikametrics — it is the honest ceiling of the entire software category, ours included where our automations run unattended. The difference between vendors is not whether the ceiling exists. It is who is responsible for the work above it, and whether that person is on your payroll or theirs.

How to read the case studies, including ours

Teikametrics publishes customer results, including a Funko case study headlined with a +131% increase in total purchases from a full-funnel strategy using Amazon Marketing Cloud. That is a good outcome and there is no reason to doubt it happened.

It is also a number with four missing pieces, and asking for them is not scepticism — it is the standard diligence any vendor should expect, and we would want you to apply it to us word for word.

• What is the denominator? Total purchases across the account, or within the campaigns that changed? A large percentage on a small base is a different result from the same percentage on the whole business.
• Over what period, against what baseline? Year on year, or against the preceding quarter? Seasonal categories can produce three-figure movements without anyone doing anything.
• What is the counterfactual? What would have happened without the change — was there a holdout, a geo split, an incrementality test, or is this a before-and-after?
• Who computed it? If the party being measured designed and ran the measurement, that is worth knowing. It does not make the number wrong; it makes it a number that should be reproducible on your own account.

Apply the same four to us. When we say that across 30 advertisers in July 2026 our Amazon DSP book ran at 6.04x return on ad spend with a $5.49 cost per acquisition, the honest caveats are that it is a portfolio figure from one month, scoped to those advertisers, computed by us. The useful version of any vendor claim is the one you can write into a contract as a measurement method rather than a promise — and asking for the method in the statement of work is a fair request to make of every party in this market.

The pricing detail that changes who this review is for

The published rate is easy to find and easy to misapply. Read from Teikametrics' own pricing page on 20 August 2026: Essentials is $179/month on the monthly tab and $149/month on the annual tab, and it is scoped to accounts spending up to $10K a month on ads. Advanced and Enterprise are described as custom pricing plus 3% of ad spend over $10K, for accounts over $10K and over $100K a month respectively.

Two consequences follow that most write-ups miss.

First, the only published base price belongs to the only tier that cannot trigger the percentage. If you spend more than $10K a month — which is most people reading a review of a platform like this — the figure in the headline does not apply to you, and the figure that does apply is not published. That is a legitimate commercial choice and it is stated plainly on their page; it just means the number in circulation is describing a smaller buyer than the typical reader.

Second, a flat fee is not a flat rate. Expressed as a share of spend within Essentials' own eligibility window, $149 a month on the annual tab works out at 1.49% of spend at $10,000 a month, 2.98% at $5,000, and 7.45% at $2,000. The tier is proportionally most expensive for the smallest accounts on it — the opposite of the intuition that flat fees favour small spenders. Worth knowing before you assume a subscription is automatically the cheaper structure than a percentage.

Whatever the tier, the question to put on the call is the same one we would want asked of us: what is the base fee at my spend level, what exactly is the percentage charged on, and is there a ceiling.

Read the contract, not the feature grid

For a multi-marketplace platform the contract questions are slightly different from the usual list, because the thing that changes over a year is not your spend — it is your channel mix.

• Channel count and how it is priced. If you add TikTok Shop in month five, does the fee change, and by how much? Get the answer before you need it.
• What the percentage is charged on. Total spend across every marketplace, or only Amazon? On a genuinely multi-channel account this is a large difference.
• Mid-term tier moves. Crossing $10K of monthly spend moves you between tiers. Ask what happens on the invoice in the month it happens, and whether it reverses if spend falls back.
• Term, renewal and notice. An annual tab implies an annual commitment. Ask how it renews and how much notice leaving requires, then put that date in a calendar the week you sign.
• Written notice of fee changes — a reasonable ask of any vendor, us included.
• Data portability. Bid history, negative keyword lists and campaign structure across three marketplaces represent a great deal of accumulated work. Confirm what leaves with you.

For a sense of how heavy these clauses get elsewhere, Quartile publishes its terms openly: a one-year initial term renewing automatically for successive one-year periods, non-renewal requiring written notice at least sixty days before expiry, fees due in advance and described as non-cancellable and non-refundable, with in-account termination available only during an initial sixty-day evaluation window. That is a conventional enterprise contract and putting it in public is to their credit — but almost no comparison page in this category quotes any vendor's terms at all, which is odd given they decide more of your risk than the feature list does.

Who should buy it, and where we lose

Buy Teikametrics ARI if a meaningful share of your revenue sits on Walmart or TikTok Shop alongside Amazon and you want one platform and one reporting model across them; if you have someone whose job includes operating it weekly; and if your spend is stable enough that the tier structure will not whipsaw. For a genuinely three-marketplace brand it is one of the more coherent options available.

Look elsewhere if Amazon is effectively all of your business, in which case you would be paying for coverage you cannot use. And if nobody in your organisation will open the platform on a normal Tuesday, no bidding software of any brand is the right purchase — the failure will not be the algorithm's.

Where we lose, said plainly. Dr. PPC is Amazon-only. We do not run Walmart and we do not run TikTok Shop, so on multi-marketplace coverage this is not a close contest and you should take us off the list. And for a seller under $10K of monthly ad spend, Essentials at $149/month on the annual tab is a far smaller invoice than ours will ever be.

Where we would put our case is narrower and honest: we charge $300 a month plus 3% of ad spend, capped at $2,500 a month, month-to-month, first 30 days free, with Orbit included — and the daily operating hours are inside that fee rather than added to it. An AI agent does the work and operators from a $500M+ Amazon team supervise it. Against a subscription you operate yourself, that is a bigger invoice and a smaller workload. Which one is cheaper depends entirely on what an hour of your team's attention costs, and that is a number only you hold.

Two places to go next if they fit. If your comparison is really about how these fee structures behave at your spend, our breakdown of Teikametrics pricing runs the arithmetic with a live calculator. And if the reason your advertising looks inefficient is stock and margin rather than bidding, Dr. Stock is the product for that problem.

Dr. PPC, Dr. Stock and Orbit are products of Full Circle, a full-service Amazon management company with more than $500M in managed revenue across 100+ brands.