Skip to content

The method

How GripMinds works

A short, calibrated test. A taxonomy of 89 patterns. A multi-language report. Twelve minutes from curiosity to a map you will return to — and below is exactly how every minute of those twelve is spent.

01Open
02Calibrate
03Rank
04Score
05Assemble
06Return

01 · Open

Open your access link

You arrive on a quiet landing page. No account is required, and we never ask for your name or email. Your session is opened with a random access token. To keep the assessment honest, we do record three technical signals — your IP address, user-agent, and a device fingerprint — strictly for integrity checks. Under GDPR these are pseudonymous personal data; we treat them as such, and never sell them, share them with third parties, or use them for marketing or model training.

Why this matters

Most assessments demand an account before they let you see anything. That filter alone biases the population toward the patient, the trusting, and the institutionally comfortable. We refuse it.

02 · Calibrate

Calibrate to your reading rhythm

A short warm-up measures your personal reading rhythm and then blends it with anonymised data from many other respondents and a fairness model. The per-question clock comes out of that mix — so deliberate stalling does not buy you more time, unusually slow timings carry less weight, but careful readers are still never penalised against an arbitrary average.

Why this matters

A fixed 30-second clock punishes anyone whose first language is not the test language, anyone reading on a phone, anyone with dyslexia, and anyone who simply prefers to read carefully. Blending your personal rhythm with anonymised population data — instead of using either alone — is the single biggest difference between a fair test and one that is either gamble-able or punitive.

03 · Rank

Rank forced-choice answers

Each item is a small scene: a context (e.g. "in a meeting", "after a failure"), a question, and five answers you rank from most-like-you to least-like-you. Every answer is silently tagged with a thinking pattern. The act of ranking is the measurement. There is no obvious "good" answer to favour — every option is a legitimate way to think.

Why this matters

Likert-style tests ("How much do you agree?") collapse signal because everyone wants to look good. Forced-choice ranking softens the social-desirability problem: you must say *which* answer is more like you, and *which* is less. The pattern beneath that choice is harder to fake — the test is designed to surface your instinctive response rather than the one you would like to project. Determined gaming is always possible; this method just raises the cost of it considerably.

04 · Score

Score across every life context

On submit, we score each of the 89 thinking patterns separately for every life context we measure. The result is not a single label — it is a textured portrait, with the same trait sometimes reading strongly in one context and quietly in another.

Why this matters

A single number averaged across all of life flattens the most interesting thing about a person — the way the same trait shifts between work, learning, pressure and home. We refuse to lose that signal.

05 · Assemble

Assemble the report from the library

Your report is composed paragraph by paragraph from our curated, human-edited library — tailored to your results and written in the language you took the test in. Every paragraph was authored across all supported languages side by side, so each version reads as carefully as the next.

Why this matters

No AI generated your report. Every paragraph was written, edited, and approved by a human before any respondent saw it. The library is multi-language on principle, not as an afterthought — every language version is equally well written, because they were edited side by side against each other.

06 · Return

Read, share, return

The report is yours forever. Read it in one sitting or revisit it across months. Share it with a coach or therapist. Compare it to a re-take in twelve months — each report is anchored to the version of our content live at the time you took it, so comparisons across years stay meaningful.

Why this matters

Self-knowledge is rarely a single-event purchase. The most useful reports are the ones people return to. We design for that — no expiring login, no rented insight, no premium tier that locks the conclusions behind a paywall.

Designed to be fair

Four fairness pillars

Calibrated to your pace

The per-question clock blends your personal reading rhythm with anonymised population data and a fairness model. Slow readers are never penalised, and deliberate stalling earns no extra time.

Per-reader fairness clock

No right or wrong

Items are forced-choice between equally legitimate answers — every option is a valid way to think. We map preference; we do not grade it.

Preference, not judgement

Context-honest

You are not the same person in a board meeting and during a personal crisis. We measure each life context separately, on its own scale, and never average them into one comforting number.

Per-context scoring

Multi-language by design

Every item, answer and report paragraph is authored across all supported languages side by side — not run through a machine translator after the fact.

Native quality

Integrity by construction

Honest signal, by design

The goal is not surveillance. The goal is to make the signal honest — for you and for whoever reads the report afterwards.

Focus-loss detection

Switching tabs or windows mid-question pauses the timer and opens a modal. Repeated events flag the session for human review before any report is issued.

Tab-switch aware

Skip and timeout caps

Hard caps on skipped and timed-out items. Trying to bypass the test triggers an invalid-session flag, and a report cannot be issued for that session.

Hard limits

Internal consistency checks

Our integrity checks surface answer patterns that look inconsistent with honest responses. The exact heuristics are proprietary; when a session is flagged, the report carries a transparency note and the operator who issued the access key is informed.

Probabilistic signal, not verdict

Privacy by architecture

Pseudonymous by architecture — honest about what we record.

We never ask for your name, email or contact details. Each session opens under a random access token. We do record three technical signals — IP address, user-agent, and a device fingerprint — strictly for assessment integrity; under GDPR these are pseudonymous personal data and we treat them as such: never sold, never shared with third parties, never used for marketing or model training. Full details on the Privacy page.

Random access tokens — never names
Pseudonymous record, treated under GDPR
No account required, no profile created
Answers used only to compute your scores
No ad-tracking, no third-party pixels
No data sales, ever
Not used to train external AI models
EU-based, GDPR-compliant

Ready when you are

Twelve minutes from here to your report.

No account. No subscription. A multi-language report, yours to keep.

We use a small set of cookies

They keep the site working, remember your language, and count how many people visit. No ad tracking, no data sales. You decide.