Evaluation method
How we evaluate platforms and projects
Driven by experts, executed by AI. The same rules for every platform and project, whether it pays us or not. Here is what we check, how the scores work and what they cannot tell you.
How it works
- Step 1Experts write the rulesOne rule set for each kind of investment.
- Step 2We collect the dataPlatform statistics, project documents, licence registers, reviews.
- Step 3AI applies the rulesTo every platform and project, in 8 languages.
- Step 4People check and publishRankings and reviews are checked before they go live.
Five checks for every platform
- 1
Licence
Is the platform licensed, and by whom? Under the EU crowdfunding licence, a national licence, or none.
Warning sign: No licence, or a licence for something else.
- 2
Returns after losses
What investors actually got after defaults, delays and recoveries, next to the rate the platform advertises.
Warning sign: Only the advertised rate is published.
- 3
How your money is secured
Mortgage, pledge, guarantee or buy-back, and how strong the party behind it is.
Warning sign: A buy-back from a lender with weak finances.
- 4
Fees
Fees to invest, to hold, to sell early and to withdraw, and how they change what you keep.
Warning sign: Fees that are hard to find.
- 5
Openness about problems
Whether late and defaulted projects are reported, explained and followed up.
Warning sign: Statistics stop being updated.
Risk levels
Each platform gets one of five risk levels from the five checks. More dots mean more risk.
- LowestStrong security, a long clean record, few late projects.
- LowGood security; losses are small and well explained.
- MediumThe security or the record has gaps; some losses.
- HighLittle security, or late projects are rising.
- HighestStartups and other bets where most fail.
Risk for the return
On lists we also show how much return you get for that risk. More dots mean a better balance.
Very goodGoodMediumBadVery bad
AI score for projects
72AI · HighWould AI invest? One number from 0 to 100 for each project, from its documents and the rules for its kind of investment.
- AI 70–100 HighStrong on most checks; read the weak points anyway.
- AI 50–69 MediumMixed: some checks pass, some do not. Read the risks.
- AI 0–49 LowSerious weak points. We would not invest on these terms.
What the score looks at
- Who is borrowing or raising money, and their record
- How the money is secured
- The terms: rate or valuation, length, fees
- The platform's own risk level
- How complete the documents are
Coming: the score broken down into these five parts on every project.
Information, not personal investment advice. The AI score rates the documents the platform publishes, can be wrong, and does not predict returns.
How we rank lists
- 1Risk for the returnOur risk-and-return level comes first.
- 2Return investors gotThen the realistic return, not the highest advertised rate.
- 3Star ratingShown where users left reviews; it does not change the order.
Lists marked “best” are our opinion, based on this method. Platforms cannot pay to move up.
Where the data comes from
- Platform statistics
- Default, delay and recovery figures as each platform publishes them, in its own definitions.
- Project documents
- Key information sheets, valuations, loan agreements and financial statements.
- Public registers
- Regulators' licence lists and warning lists, company registers.
- Investors
- Reviews and reports from people who invested, checked before they count.
Limits and corrections
What the method cannot do
- Predict the future: past returns do not predict future returns.
- See what a platform hides: we can only rate what is published or sent to us.
- Replace advice: this is information, not personal investment advice.
- Be free of errors: AI and people make mistakes. Tell us and we fix them.
Spotted a mistake?
Readers and platforms can report wrong data on any page. We check every report and fix what is wrong.
Report a mistakeMethod version 1.0 · first published with the new site
Questions people ask
Do platforms pay to be listed?
No. Listing is free and every platform goes through the same checks. Some platforms pay a referral fee when you sign up through our link; those links are marked “Referral link”.
How often are scores updated?
When a platform publishes new statistics or we find new information, and at least once a year for every platform.
Who are the experts?
People who work in each field: credit, real estate finance, energy projects, startups, regulation. The About page shows who is involved.
Can a platform dispute its score?
Yes. A platform can send us data or corrections; we check them and change the score when the evidence supports it.