What Kinetry cannot tell you
Every measurement method has limits. Kinetry's are listed here, at full strength, next to what the product does about each.
A Kinetry score is a record of how the people around someone experienced their behavior — filtered through attention, mood, generosity, and everything else human raters carry. It is not a physical measurement, and two honest raters can genuinely disagree.
The fairness rules exist for exactly this: rater minimums in 360 mode before individual scores show, “not observed” as a real answer (up to 6 of the 24 per assessment), and rating-integrity checks that nudge raters off autopilot. Structure reduces the noise; nothing eliminates it.
The Kept-Commitment Method establishes ordering — commit first, measure after — and reports the contrast between kept and not-kept commitments. That is a correlational, observational comparison over real work, not a controlled trial, and people who keep commitments differ from people who don't in more ways than the commitment.
Every surface that shows the numbers says “preceded,” never “caused,” and the full procedure, floors, and limitations are published on the method's own page.
Inside one workspace, a behavior's score can rest on a handful of ratings, and aggregate contrasts can sit below the floors where a number would be noise, not signal.
Below the floors, the aggregate surfaces stay silent rather than print a number — which means Kinetry will sometimes tell you nothing at all. Your own private “Since last cycle” ledger is the stated exception: a personal record, not a statistic. An unmeasured behavior is never treated as a zero.
Kinetry measures observed workplace behavior, cycle by cycle. It has no personality model, no trait types, no cognitive assessment, and no hiring-fit prediction — and behavior scores must not be read as statements about character or potential.
That's a boundary, not a roadmap gap: trait instruments answer different questions, some of them well. Kinetry's premise is narrower — what people do can change, and change is what it measures.
Kinetry is a coaching and development instrument, not an employee-evaluation system. It has never been validated for employment decisions, and using it as the sole or primary basis for pay, promotion, or termination is against its Terms of Service.
The terms say it, the product's guidance says it, and this page says it. If you need a legally validated selection or evaluation instrument, this is not one.
Nothing in Kinetry assesses anyone's psychological state. The rating flow's mood check exists to protect score quality from the rater's day, not to evaluate the rater.
Behavioral data stays behavioral. If a team needs wellbeing support, that's a different tool and often a different profession.
When one outside advisor is the sole scorer, the scores carry that advisor's lens — consistent and experienced, but singular. Single-advisor ratings also land on the anchor points of the rating scale, so advisor-mode numbers sit on a different footing than 360 numbers — Kinetry never pools the two, and neither should you. It is not a 360 consensus, and it doesn't claim to be.
The mode exists because a consistent expert lens applied every cycle beats no measurement at all — and score surfaces state which assessment mode the workspace runs.
A 74 in one workspace and a 74 in another are not the same fact: different raters, different severity, different context. Kinetry publishes no cross-company benchmarks today, and workspace scores are calibrated only within the workspace.
Benchmarking machinery exists behind participation floors and stays off until enough companies join to make a comparison honest — being included is optional either way.
A fixed instrument is what makes movement comparable cycle after cycle — and it means everything outside the 24 goes unmeasured: domain skill, output quality, results, judgment calls the definitions don't cover.
The definitions are public, so you can see exactly what is and isn't in scope before trusting a number with anything.
Where AI coaching is enabled, the prose is generated — and generated text can emphasize oddly or read more confidently than the underlying numbers deserve.
The AI never computes or alters a score. It writes from the same computed facts the page shows, cites its evidence, and is off by default until a workspace turns it on.
Within those limits: how the people around each person — or their advisor — experienced 24 defined behaviors this cycle, how that compares to every previous cycle on the same yardstick, which pairs share strengths and gaps under the Kinetry Pair Model, and whether the behaviors people committed to work on moved after they kept the commitment — the Kept-Commitment Method. The fairness machinery is at How scores are protected, the instrument at The 24 behaviors.