How we test

How a tool ends up on our list

We don't test everything that exists. We test what is genuinely relevant for e-commerce founders in German-speaking Europe. A tool makes the list if at least one of these applies:

• It solves a problem readers keep telling us about, or one we run into in our own business.
• It is actually usable in the DACH region: German interface, EU hosting, GDPR-ready contracts or German-speaking support.
• It shows up in our audience, in forums, communities and search queries, and there is no useful independent write-up yet.
• It is affordable for beginners or offers a meaningful free tier.

We do not charge for a tool to be tested and we do not sell review slots. Vendor requests are treated like any other suggestion. They can get a tool onto our list, but they do not influence what ends up in the article.

The scoring framework

A rated tool is scored across five dimensions, each from 0 to 5 points. We adapt the labels to the tool category: a search product gets “Search quality & AI features”, an accounting tool gets “Feature set & DATEV”. When one area matters especially for the buying decision, it gets its own sixth score, such as “Contract & cancellation” for sevDesk or “Legal compliance in DACH” for Shopify. Behind the labels, the same five questions always apply:

How the overall rating is calculated

The overall rating is the unweighted average of the individual scores, rounded half up to one decimal place. There is no hidden point system and no weighting we don't show you. You can recalculate every rating that comes with a score box, here using our Doofinder review as an example:

Next to every overall rating we add one sentence explaining where the points were lost. If you're short on time, that is the most useful sentence in the whole review, because it tells you what will annoy you about the tool.

Since October 2026 we check every overall rating against its individual scores with a script. The first run flagged 16 German and 14 English reviews where something was off in the review header or in the rounding. What we corrected is listed in the corrections log below. All current ratings are collected in our tool finder.

Examples of explained deductions

This is what the reasoning behind a low individual score looks like. The table shows four examples from current reviews, with check date and sources.

Where our data comes from

We can't run every tool in live production for months. So we're explicit about what each rating is based on. It's always a mix of these sources:

• Our own test. Create an account, set it up, run the core use case on a real case. This is the basis for the usability, setup and feature-set scores.
• Documented hands-on tests by others. Where we have no access of our own, or a published test shows the workflow in more detail, we use it and name it as a source. Example: the walkthrough in our UX Pilot review follows a video test by Nick Babich.
• Vendor information. Pricing pages, documentation, terms of service, privacy policy, support hours. We take these as vendor claims and label them as such. We read pricing pages in the browser, not in the source code, because some vendors swap prices through A/B tests. That was the case with sevDesk in October 2026.
• Independent review platforms. Trustpilot, G2, Capterra, OMR Reviews, the app stores of the shop systems, plus reports from forums and Reddit. These mainly feed “Support & track record”, because a single test says little about that. We always quote these ratings with count and date, and they are never our own score.

Where our tests hit their limits

Four constraints we won't argue away, so you can judge the result properly:

• We typically test for days to weeks, not years. Long-term issues such as creeping price increases or declining support often only reach us through third-party sources.
• We test from the perspective of a small to mid-sized shop. Whether a tool still holds up at six-figure order volumes is usually beyond what we can verify.
• Software changes. Every review carries a last-updated date, and whatever happened after that only appears with the next update.

Changes and corrections

When we change a review substantially or correct a mistake, we record it here. The table shows the most recent entries.

Independence & affiliate links

We partly finance this site through affiliate links. If you buy through one of them, we earn a commission at no extra cost to you. To keep that verifiable, these rules apply:

• Affiliate links are marked with an asterisk (*) and a note in the article.
• Commission rates are not a criterion in the scoring framework. They simply don't appear in it.
• We also review tools we earn nothing from, such as Shopify, Brevo and Printful, and we write down weaknesses even when there's a partner programme behind the tool.
• No vendor gets to approve an article before publication.

We're not claiming this makes us automatically neutral, because a business model always exerts pressure. But you can check whether we stick to it. Most tools we review have a partner programme, including those with the weakest individual scores: Satellite with 2.5 for “SMS & emergency calls”, Higgsfield with 2.8 for billing transparency and support, and DiscountHero with 3.0 for track record. For Funnelforms we removed our affiliate links in August 2026 because we currently don't recommend the tool.

Frequently asked questions about our tool reviews

Found a mistake?

If a price is out of date, a feature has disappeared, or your day-to-day experience with a tool differs from ours: tell us. We correct reviews and reset the last-updated date. That happens regularly, and we much prefer it to an article quietly going stale.

Report a correction

Who does the testing

This page was last reviewed on October 15, 2026. We update it whenever our process changes or we correct a review.