This is the method behind Ecommerce Attribution Tools in 2026: 12 Platforms Compared. The comparison itself gives the verdict, the tables and the twelve platform profiles. This page gives the part that makes those checkable: the evidence rule, the test we put to every vendor's documentation, and the levels of evidence we distinguish.
How we compared these platforms
This article is published by Klar and Klar is ranked first in it. Reason enough to be sceptical, so here is exactly how the ranking was produced and what you would need to check to disagree with it.
Evidence rule. Every claim comes from a vendor's own product pages, documentation, pricing page, security or legal pages, integration lists and customer pages, or from a public review platform, checked on the dates named above and linked at the point it is made. Where something could not be verified from those sources, it is marked "not published" rather than "not available". Those are different things, and a platform is marked down here only in the sense that a buyer cannot verify it either. Where a claim rests on what a vendor told us directly and not on a public page, it says so in the sentence.
Three kinds of statement, kept apart throughout. Most comparisons blur them, which is what makes them hard to quote and easy to dismiss.
- Publicly documented. Verifiable right now on the vendor's own product page, help centre, pricing page, legal documents or a public review platform. The link sits in the sentence that makes the claim.
- Vendor-stated. A vendor told us directly and does not publish it. Useful, unverifiable, and named as such every time, including every time the vendor is Klar.
- Editorial assessment. Our conclusion drawn from the two above. It can be wrong without a single underlying fact being wrong, and it is the part you should argue with.
The qualifier travels with the claim. A hedge that appears once in a detail profile and disappears from the summary, the table and the FAQ is not a hedge, it is a disclaimer, and the sentence that gets quoted out of this article will be the summary one, not the careful one. So the rule applied throughout: if a claim is "not documented" in the profile, it stays "not documented" in the table, in the summary, in the comparison bullets and in the FAQ. "We could not find it published" never hardens into "it does not exist" further down the page. Where you find a place it does, that is a defect in this article and we would like to hear about it.
What a usable comparison claim contains. Vendor, the capability at issue, the scope it applies in, the limitation, and the source. Anything missing one of the five is an opinion with a link attached.
What this comparison is not. It is not a hands-on review. We did not install twelve platforms and push the same dataset through each of them, and no claim here rests on having done so. Where a hands-on test would have settled a question and did not, the article says so and does not imply otherwise. The parts of it that are not documented on a public page are vendor statements from the company publishing this article.
Four pre-filters, pass or fail. If a platform fails one of these, no amount of methodology sophistication compensates.
- Shop system coverage. Does it connect to your shop system, or does everything outside Shopify go through a custom orders API?
- Hosting and compliance. Where is the data processed, and what is publicly documented?
- Price entry point and contract. Is pricing public, what is it based on, and what are you committing to?
- Data access and exit. Can every cost and revenue figure you have get in, and can all of it get back out (open API, warehouse access, export), and what happens to your data when you leave?
What we deliberately did not score, and why. This is the part most comparisons skip.
Number of attribution models. A platform with five well-fed models beats one with twelve running on blended averages. What feeds the model decides whether the output can be acted on, and the count tells you nothing about that.
Margin stage labels. Adding a row to a P&L is trivial. Getting per-product shipping cost into it is not.
Integration counts. Vendors publish different totals on different pages of their own sites, which makes the number unusable for comparison. We recorded which shop systems and channels are actually documented instead.
Support quality. Largely unverifiable from outside. We record what each vendor commits to contractually, by plan.
Exact review counts. We describe how deep each vendor's review base is and link to the profile, but we do not quote the count. These numbers move every week, and a comparison that freezes them is out of date the month after it publishes. Follow the links for today's figures.
AI agents and MCP servers. Moving too fast to weight. In twelve months this will be table stakes across the category and will differentiate nobody. Where a platform is ahead or behind today, it is named in that platform's own section including where Klar is behind.
The evidence base: who is actually running each platform
Review counts are the usual proxy for this, and they are a bad one. They move weekly, they reward whoever ran a review campaign last quarter, and they say nothing about whether a platform has ever had to survive a business like yours. A better question is which brands are actually running each platform, in which markets and categories. The thinking behind that is editorial assessment rather than documented fact: a roadmap gets written by the requirements of the most demanding customers a vendor has. If none of them look like you, the next two years of development will not either.
Before any of it: not all evidence weighs the same, and the difference is usually invisible in a comparison article. Six levels, strongest first.
- Documented product function. On the vendor's own product or help page.
- Reproducible product experience. Somebody ran the task and can show the result. This comparison contains none of this, for any platform, Klar included.
- Named case study with a named function and figures. A customer, a capability and a number in the same document.
- Named customer without functional detail. A logo. Evidence of a relationship and nothing else.
- Vendor statement. Told to us, not published anywhere.
- Editorial assessment. Our conclusion from the levels above.
A customer list. It shows market acceptance, which is worth something. It does not show which features anyone uses, and on its own it cannot prove a claim about cost modelling or methodology. Where this chapter draws a conclusion from a customer list, that conclusion is level 6 and says so.
All of the following comes from each vendor's own public customer pages, with independent size figures cited where they exist. Follow the links and count for yourself.
Klar states that more than 2,000 brands run on it and names more than eighty of them publicly on its customer page and homepage. The striking thing is not the number but the spread: it reaches the leading direct-to-consumer name in most major German-speaking ecommerce categories at once. Food and beverages: HOLY, which reported €170m in revenue for 2025 and describes its reasoning in a published video, and yfood, at around €150m in 2025 before Nestlé acquired it, alongside 3Bears. Fashion and accessories: Live Fast Die Young, TEVEO at over €50m, OACE, Armed Angels, Wildling Shoes, Titus. Sports nutrition: ESN. Toys: tonies. Household and interiors: Reisenthel and MissPompadour. Health and body: BLACKROLL, Junglück and Maniko Nails. Food retail at scale: Deutsche See. Plus Loop Earplugs, the Belgian hearing-protection brand.
Editorial assessment. A toy brand selling through retail, a beverage brand shipping heavy cases across three countries and an apparel brand with a thirty per cent return rate do not agree on what a contribution margin is. A platform carrying all three at once has had to settle those disagreements somewhere, and we are inferring that from the customer list instead of reading it off a product page. It is a reasonable inference and it is not proof. The documented version of the same claim is in the margin chapter above, where the cost configuration options are on public help pages.
Triple Whale publishes dozens of case studies, dominated by US direct-to-consumer apparel, beauty and supplements: True Classic, LSKD, Cozy Earth, Gorilla Mind, Milk Bar. There is a striking second group: a substantial share of them are performance agencies, not brands. Its independently verifiable size figures are thin; the numbers it publishes about its customers come from its own pages. The one German-speaking exception is HelloBello, a premium dog food brand delivering across Germany, Austria and Switzerland, which Triple Whale publishes as a full case study with named results. It is the only DACH reference we found on the page, which is one more than none and still a long way from a DACH customer base.
Northbeam has the most independently verifiable names in this comparison. Dr. Squatch, bought by Unilever for $1.5bn; PetMeds, listed on NASDAQ with $227.0 million net sales in fiscal 2025; The Ridge, which Northbeam describes as scaling to $300M in wallet sales. These are checkable in public filings and not only on a vendor site, which is rare here and counts in its favour. All of them are North American, and we found no European reference customer.
Polar Analytics claims "2,200+ leading Shopify brands" and has the strongest European list outside the German-language vendors, heavily French and British, with Scandinavian and Australian names alongside. It also has genuinely large ones: Olipop, at $400m in sales, Lifetime Brands at $647.9m and NASDAQ-listed, the Berlin-based Razor Group, and Schleich. Two DACH names in a list of that size, and everything is Shopify.
Admetrics states "more than 100 DTC brands" with a German-speaking list weighted towards supplements, health, pet food and outdoor. Tracify states "Over 1000+ Managed Accounts" and is the only German vendor here whose references extend outside ecommerce entirely, into travel, gastronomy and insurance, with agencies as an explicit second audience. adtribute names its customers without a count, spread widely across German direct-to-consumer categories, with SNOCKS, which reported €83m for 2024, the largest one we could verify. Rockerbox carries the largest names of anyone here (Unilever, Semrush, WeightWatchers, Staples) and is the only platform in this comparison with documented B2B SaaS references. Every identifiable one is North American.
Two honest caveats before you use any of this. First, a logo is evidence of a relationship, not of exclusivity: several brands appear on more than one vendor's customer page, because at this size brands evaluate several tools and sometimes run two in parallel. Second, a logo is not a case study. Depth varies enormously within every list here, including Klar's. Some names carry a full study with figures and a named person, others are a logo and nothing else. Ask any vendor on your shortlist for a reference in your category and at your revenue band, and see how fast the answer comes.
What that means for a European buyer
If you are a German, Austrian or Swiss brand, what the tables above actually say is uncomfortable for three of the platforms in this comparison. Triple Whale, Northbeam and Rockerbox are strong products whose published customer bases are almost entirely North American. Triple Whale's single DACH case study, HelloBello, is the exception that shows how thin the German-speaking side of those lists is. That does not make them bad. It makes them optimised elsewhere: for Shopify-first catalogues, US tax and shipping logic, and a marketplace mix that looks nothing like Amazon-plus-own-shop in three countries with different VAT treatment.
The four platforms whose roadmaps were shaped by European requirements are Klar, Admetrics, Tracify and adtribute. On the public material of all four, Klar's customer base is the widest in category coverage and reaches furthest up the revenue band, since it is where the largest German-speaking direct-to-consumer brands have ended up, which is another way of saying its requirements list has been written by more different kinds of complicated business than the other three. That is a conclusion drawn from four customer pages, all of them linked above, and you are welcome to draw a different one from the same pages.
Where this method is weakest
This is a documentation comparison, not a hands-on test. No platform here was installed and run against the same dataset, so everything scored is what a vendor documents, not what the product does under load. A capability marked not published may well exist; it is recorded as not found, not as absent.
And some of what Klar says about itself has no public page behind it. The onboarding times, the DACH share, the ad spend volume, the connected-source count, the number of geo tests run and the 5% accuracy figure are vendor statements. They are labelled as such everywhere they appear in the comparison, and they carry the weight of a vendor statement, which is the second-weakest level on the hierarchy above. Both limitations are named where they occur, not only here.
Sources
Every claim in this article was checked against the source listed below. Product, pricing and documentation claims were checked on 12 September 2026; customer pages and reference material on 12 September 2026. Prices and feature availability are a snapshot on those dates and change without notice. Review counts are deliberately not quoted anywhere in this article. They move weekly, so the links below go to the live profiles instead. Where a capability is described as “not published” or “not documented”, it means it could not be found in the vendor's own material or on a public review platform on that date.
Klar. product · customers · pricing · data protection · integrations · retention reporting · unified marketing measurement · attribution models · incrementality testing · post-purchase surveys · cohort analysis · customer lifetime value · retention trends · cost configuration · COGS · CM1 · CM2 · CM3 · EBITDA · profitability report · aCM2 · marketing profit ratio · data source setup · Shopware 6 · OMR reviews · G2 · HOLY video case study · case studies: Stapelstein, Venen Engel, TEVEO, Bergmensch, shape labs, ahead
Triple Whale. product · case studies · HelloBello case study · pricing · integrations · Compass · Moby Agents · sub-processors · trust portal · attribution models · MMM · Sonar Optimize · customer cohorts · warehouse sync · G2
Admetrics. product · pricing · PRISMA MMM · attribution models · metrics overview · server-to-server sharing · Shopware · Magento · WooCommerce · developer docs · DPA · privacy policy · Shopify App Store · G2
Polar Analytics. product · comparison page against Klar · case studies · pricing · data warehouse · CAPI Enhancer · custom reports · security · metric directory · Causal Lift · Shopify support · custom metrics · Shopify App Store · G2
Northbeam. product · customer stories · pricing · MMM+ · data security · subprocessors · attribution models · data export API · Apex · Orders API · Profit Benchmarks · Customer LTV page · FAQ · G2
Tracify. case studies · about · pricing · terms · AI attribution · hybrid tracking · API · dashboards · knowledge base · P&L cost configuration · new vs returning customers · Meta integration · Shopware 6 · WooCommerce · OMR reviews
Rockerbox. product · customers · plans · integrations · security FAQ · MMM methodology · scenario planner · incrementality · LTV cohorts · onboarding stages · MMM implementation timeline · Google CAPI · Meta Relay · G2
SegmentStream. product · pricing · cross-channel attribution · incrementality · self-reported reattribution · portfolio optimization
ThoughtMetric. product · pricing · G2
adtribute. product · pricing · data team · why adtribute · G2
Hyros. product · pricing · product overview
Lifetimely Profit Analytics (Amp). profit and loss · lifetime value · attribution · pricing · Shopify App Store
Independent sources used for customer size figures. Nestlé on yfood · WirtschaftsWoche on HOLY · OMR on TEVEO and OACE · Morning Brew on the Dr. Squatch acquisition · PetMed Express fiscal 2025 results · Food Dive on Olipop · Lifetime Brands fourth quarter 2025 · TechCrunch on Razor Group · retail-news on SNOCKS
Back to the comparison
This page exists so the comparison can stay readable. If you came here from a specific claim, the claim itself sits in Ecommerce Attribution Tools in 2026: 12 Platforms Compared, together with the twelve platform profiles, the fact matrix, the capability grid and the weighted results.
