See yourself as your team sees you — then measure whether it changed.

Two ways that do not depend on anyone behaving well. Responses from anonymous rater groups are stored without a rater identity — a database trigger reads the cycle's own instrument settings and drops the identity at write time — and no anonymous group with fewer than three completed responses (the default threshold, which cannot be set lower) is ever shown on its own; it is merged with the other groups first. Open comments are AI-anonymised and shuffled, and withheld entirely if that step fails.

Self versus others for every competency in the instrument, rater-group breakdowns where the threshold allows, a quadrant placement for each competency — confirmed strength, hidden strength, blind spot or development area — anonymised open comments, a development focus of up to three competencies, and, when a previous cycle exists, the change against it. When importance rating is on it also ranks the gaps that matter most for the role. The subject sees the same report the administrator released, on a private link or as a PDF.

The composer builds an instrument from the 83-item bank against the 29 competencies of the AssessAll Capability Framework, and four templates ship ready for leaders, managers, sales professionals and individual contributors. Organisations mapping their own framework to the AACF can run cycles on that mapping; fully bespoke item writing is available through the custom-assessment service.

A follow-up cycle is created from a closed one: it copies the subjects and raters with fresh personal links and freezes the same instrument snapshot, so both cycles measure identically. Each subject's second report shows the movement per competency against the first, and the cohort page shows the change across everyone — the evidence a leadership programme needs.

Fifty credits per subject by default, charged only when a report is released — ₹1,500 or US$25 at the standard credit rate, priced in INR and USD, with client-specific pricing available. Rater invitations, responses and draft reports are free until release. Raters and subjects need no accounts, so cycles run wherever the people are.

Three is the working floor, and the arithmetic says so twice over. AssessAll enforces an anonymity threshold that defaults to three and cannot be configured below it: an anonymous group with fewer completed responses is merged into the combined others view rather than shown alone, and if the merged view is still short the scores are withheld — with n=2 a distribution is simply the individual answers. Separately, the reliability of an averaged rating follows Spearman-Brown. At the interrater reliability of 0.45 that the published literature reports for ratings collected for administrative purposes, one rater is 0.45, two are 0.62, three are 0.71 and five are 0.80 — so the third rater is the last one that buys a large improvement, and a floor built for anonymity lands on roughly the right number for evidence. One correction, because that 0.45 has since been superseded: Zhou, Sackett, Shen and Beatty (Journal of Applied Psychology, 2024; 132 independent samples) report supervisory interrater reliability at 0.65 overall and advise against using any single grand mean, because it varies by the job being rated — 0.57 for managerial positions against 0.68 for non-managerial ones. A 360 cycle is usually run on managers, and at 0.57 two raters reach 0.73, three reach 0.799 and it takes four to pass 0.80. Aim for four or five per anonymous group: that is what the newer figure requires for evidence, not only what survives one non-response.

Nothing, deliberately — and that is the point. A “not observed” response is excluded from every average rather than treated as a low rating, because a rater who has never seen the behaviour is not a rater who thinks it is done badly. Each item and competency therefore carries an observed count alongside its average, so a low-evidence number is visibly low-evidence. Read the two together: a low score with a high observed count is a development need, and a low score with an observed count of two is a visibility problem, which is a different conversation and a different intervention.

Yes, if the programme team turns it on. Each subject gets a personal link to propose raters inside every group's minimum and maximum. Their proposals are not invited until they are approved — by the subject's named manager through a separate approval link, or by the programme team — and the approver can remove anyone who does not see enough of the work. The approver sees who is on the list, never what anyone answers, and anonymous groups are still only ever reported combined.