DatascaleIntegrationsMethodology
Review methodology
How we rate. And what we earn on.
Every score in the catalog is an editorial verdict from our project work. Where a partnership exists, it sits right on the rating, not in the fine print.
The criteria
How the number is made.
A 3 is solid and production-ready; a 5 is the reference in its role. There are no courtesy fives: real weaknesses pull real criteria down.
Each tool gets three to four criteria, all phrased so higher is better. The set is fixed per category and identical for every tool in it, otherwise the comparison matrix lines up unequal rows. The full list is below.
The mean of the criteria visible on the page. Nothing added, nothing hidden; a deviating weight would sit visibly next to it.
EU data sovereignty is its own criterion, not a footnote. US tools honestly lose points there and stay recommended where the job needs them.
The sets
Which criteria apply in which category.
Every tool in a category is rated on the same axes, in the same order. This table is generated from the file the build also checks against; it cannot disagree with the actual set.
| Category | Tools | Criteria |
|---|---|---|
| Analytics | 9 | EU data sovereignty · Analytical depth · Value for money · Ease of use |
| BI & Visualisation | 4 | Accessibility · EU data sovereignty · Analytical depth · Value for money |
| CDP & Event Pipelines | 4 | Ecosystem · EU data sovereignty · Value for money · Ease of use |
| Consent & CMP | 4 | Legal certainty · EU data sovereignty · Value for money · Ease of use |
| CRM | 2 | Ease of use · EU data sovereignty · Ecosystem · Value for money · Scalability |
| Data Warehouse | 3 | Scalability · EU data sovereignty · Value for money · Low maintenance |
| Data Integration & ETL | 4 | Ecosystem · EU data sovereignty · Reliability · Value for money · Ease of use |
| Marketing Automation | 2 | Activation · E-commerce fit · EU data sovereignty |
| Server-Side & Tagging | 2 | EU data sovereignty · Data quality · Value for money · Ease of use |
| Tag Management | 2 | Ecosystem · EU data sovereignty · Value for money · Ease of use |
Ad platforms, DSPs and healthcare media deliberately carry no score and are therefore absent from this table.
7 rules
So a score is worth something.
Production experience only
We rate what we run ourselves or have built into production for clients. No tool gets rated from its docs alone.
One tool, one score
Catalogue, detail page and hands-on review read the same file. There is no second place a rating could live, and therefore none where it could differ. A build check compares the shipped pages against the source and fails when a number, a criterion or a disclosure sentence drifts apart.
Score and commission stay separate
With stape.io, funnel.io, Usercentrics and OneTrust we earn on licences. Checkable rather than claimed: the four partners average 4.33, the other 33 average 4.01, and the worst score in the whole catalogue belongs to a partner at 3.40.
Partner tools name the alternative
Every rating of a partner tool carries at least one alternative we have no commercial relationship with, in the same place and the same presentation. Not as a footnote. The build check verifies it against the shipped page.
Tradeoffs are spelled out
Every rating names what you give up with the tool. A score without a list of limits does not get published.
Changes are public
Every rating carries a review date and a change log: what changed, when, and why. It sits on the tool page, not in an internal ticket. When a vendor changes prices or hosting, we re-review and add an entry rather than quietly overwriting the old number.
No user reviews
The scores are not aggregated user ratings and do not pretend to be. Site-wide, which includes the blog and every hands-on review: no stars, only criterion bars with the number beside them, and nowhere a self-issued aggregateRating in the markup. A named editorial verdict, nothing more.
All commission relationships at a glance
The full table, including the row where no partnership exists, lives on the about page.
Catalog
36 integrations, rated by these rules.
With search, categories and head-to-head comparisons. The decision for your stack is still made in the audit, with your numbers.