What do you actually get at the end?

The single biggest question about any assessment is what lands in your inbox afterwards. Here are the real artefacts — downloadable without a form, and annotated panel by panel so you can read one whether or not you have seen a report like it before.

10 questions to ask of any sample report

These work on anyone's report, ours included. Several publishers in this market post sample PDFs without a form — download a few and put the same questions to each.

1. 1

   Does it name the norm group the scores are compared against?

   A percentile with no stated comparison group is not a measurement. Look for it on the cover, not in a footnote.

2. 2

   Does any paragraph cite the reader's own answers?

   Interpretation text can be assembled from a lookup table. A line quoting which option the person actually chose cannot, and it is the clearest sign the report was generated rather than templated.

3. 3

   Is there a response-quality or validity check, and what does it do?

   A check that lowers a confidence band is a measurement instrument. A check that auto-rejects the person is a filter, and should be evaluated as one.

4. 4

   Does the critical column cost anything?

   If the blind-spots section reads as flattery in the shape of a critique, the report is a sales document. A usable one names something the reader would rather not read.

5. 5

   Is the advice specific enough to act on this week?

   'Communicate more clearly' is unusable. 'Send the numbers before the meeting, not during it' is a behaviour someone can change on Monday.

6. 6

   How many people is the comparison built from, and is that number printed?

   A percentile or a 'vs norm' delta computed from a handful of prior attempts is a running average of whoever sat first, not a norm group. If the sample size is not on the artefact, ask for it before you quote the number to anyone.

7. 7

   What does the integrity or proctoring summary print when nothing was recorded?

   Some reports default to the most favourable label when no data exists, which makes an unproctored sitting look like a clean one. Read the counters beside the badge: all-zero counters with a zero photo count is an absence of evidence, not evidence of absence.

8. 8

   Is a minimum-group or anonymity threshold stated, and does it apply to the scores, the written comments, or both?

   On a multi-rater report these are usually two different numbers applied to two different groupings — a floor per rater group for the scores, and a floor across the whole pool for the comments. The second one is what decides whether a two-person team's written feedback is really unattributable. Almost nobody publishes either; ask for both.

9. 9

   Does anything on the artefact say how much of the number is measurement error?

   A bar built from one rater and a bar built from six are printed in the same typeface to the same decimal place, and they are not the same kind of evidence. A reliability figure, a standard error or a confidence interval is what tells the two apart. Checked across nine published sample reports in September 2026, not one carried any of the three — so this is the question the whole category currently fails.

10. 10

    Does the report say what it does not measure?

    Every behavioural instrument has hard limits — no ability signal, weak links to job performance, no clinical meaning. A report that never mentions them is more likely to be over-read by whoever receives it.