Our test methodology
Every app in our test goes through the same procedure: eight fixed criteria, a uniform scale from 1 to 5 and a published calculation for the overall score. This page documents the procedure in full, so every score can be traced.
We test for the German-speaking market, with a particular eye on the Albanian community. That is why Albanian fit is one of the eight criteria and counts as much as any other. We say this openly because it shapes the overall ranking: international apps without a connection to the community lose points there, Albanian apps gain them.
The eight test criteria
1.Seriousness
Measures whether an app is built for committed relationships or for casual encounters. We assess the product's evident orientation: profile depth, matching logic and the intentions the app assumes of its users. A 1 stands for purely casual, a 5 for clearly relationship-focused.
2.Pricing fairness
Assesses how much an app actually delivers without payment and whether its pricing model is transparent. Points are deducted for paywall traps, meaning core features that only surface behind a payment barrier after matching. A fair free model earns high scores.
3.Gender balance
Captures how balanced the ratio of women to men in the user base is. A heavily skewed pool noticeably worsens match chances for the majority side, regardless of how good the app itself is.
4.Albanian fit
Assesses the fit for Albanian and Balkan-rooted dating: language, cultural understanding and whether you can actually find people with a similar background there. International apps without any cultural angle score low here. Without this criterion, Tinder, Bumble and Hinge would each score roughly a quarter to a third of a point higher overall, and Hinge would rank ahead of Dua.
5.Safety and verification
Examines how an app separates real people from fake profiles and which protection mechanisms it offers. We assess authenticity checks such as photo verification as well as reporting and blocking features.
6.Usability
Rates the everyday experience in the app: clarity of the interface, speed and how smoothly the path from discovery to chat works. Visual design only counts insofar as it genuinely helps or hinders actual use.
7.Privacy
Rates how personal data is handled: what is collected, how transparently the app communicates about it and how much control users keep. The basis is the providers' publicly available privacy documentation.
8.Reach
Measures the size of the reachable matching pool. A large user base widens the selection but says nothing about its fit; that is why reach deliberately carries the same weight as the other seven criteria.
Scale and calculation
Each criterion is scored from 1 (weak) to 5 (strong). The overall score is the unweighted average: the eight individual values are added up, divided by eight and rounded to one decimal place. No criterion counts more than any other.
An example from the current test: an app with the individual values 5, 5, 4, 5, 5, 4, 4 and 2 reaches a sum of 34; divided by eight that is 4.25, rounded to an overall score of 4.3.
The average has a deliberate consequence: apps with very different profiles can end up numerically close. So while the overall ranking sorts by points, we do not crown a winner for every situation. That is why we also name category winners.
What the ratings are based on
The scores come from three sources: our own use of the apps, an examination of their pricing and feature models, and an evaluation of publicly available reviews and information on safety and data processing.
One exception applies to announced apps that are not in the App Store yet, currently embla. There is no hands-on use there. We derive seriousness, pricing fairness and Albanian fit from the app’s rules, and verification from an announced process that nobody has been able to test yet. Reach, usability, gender balance and privacy depend on daily use with real people. That is why all eight scores are provisional, marked ‘Preview’ everywhere, and will be re-checked after launch.
Where claims cannot be verified reliably, we rate conservatively instead of speculating.
Update rhythm
Tests are updated whenever something substantial changes in an app, such as prices, core features or safety mechanisms. The date of the last update is shown on each article, and the comparison matrix carries the date of the test. The current round carries the date October 2026.
No paid placements
No provider can buy a place in the table or a better score. Links to providers may generate a commission; this has no influence on scores, ranking or verdicts. Where an app is weak, the test says so.
Corrections
If a statement turns out to be wrong or outdated, we correct the affected passage and adjust the rating. A dedicated contact channel for corrections does not exist yet; until it does, we regularly check our own statements against publicly available sources.