A concept test report is built to drive a build-or-pivot call, so it leads with a verdict per concept and works backwards to the evidence. This article covers the parts specific to concept testing. For the sections every report shares, plus exports, judging how much weight a finding carries, and the thin-run banner, start with Reading a synthesis report.
Up to four callouts can appear near the top of What Matters Most, just under the headline. Each one is telling you that a result further down needs reading differently.
The first three link straight to the concept they refer to, further down in Findings. The audience-fit one doesn’t link. When the report can tell which concept the group pulled toward, the box names that concept, and a Drop on it means don’t advance it with the audience you tested. If any of these boxes is present, resolve it before you act on a Drop.
Also in What Matters Most: one row per concept, ordered best to worst, each with its verdict, how many participants named it their favourite (for example 18 of 24). The favourite count appears there and under the concept’s Intent score, nowhere else. When the concept with the most favourites is not the one to pursue, the headline box says so in a sentence of its own, and the next sentence gives the reason it is not the one to pursue now. A concept with more favourites never shows a lower Intent grade than one with fewer. When the grade was raised to match the favourite count, the count sentence under it is the whole explanation. The next step for each concept sits in Next Steps, and the concept’s own row further down in Findings links to it. If you read nothing else, read this. It is the whole study compressed into a table you can take into a roadmap conversation.
This is the single most misread thing in any Candor report, so it’s worth stating plainly. Severity describes how hard an objection is to answer, not how strongly participants voiced it.
So a wall of Low-severity objections is a healthy concept with a messaging job to do. A couple of High-severity ones is a concept that needs rethinking, however enthusiastic the rest of the report sounds.
The concepts one by one, sorted best to worst by verdict rather than by the order you entered them. Each row carries the verdict, a summary, the description, and a collapsed detail with the evidence. On a card-sort test (five or more concepts) the row also says how many participants were probed in depth on that concept.
Alongside them sit direct answers to the learning goals you set, and a section naming needs that none of your tested concepts covered. That last one is the closest a concept test gets to discovery, and it’s where the next idea usually comes from.
The action for each concept appears once, in Next Steps, rather than being repeated under every concept. A concept row that a step covers carries a See Step N link; clicking it opens Next Steps at the step. A concept no step names carries none.
Three things render only when the run used comparative mode, which means you tested two to four concepts: the head-to-head read on how participants chose, What would have to be true to adopt, and, for B2B audiences, Approval requirements covering who else has to sign off.
The middle one is the most practical section in the report, because it converts objections into conditions you could go and satisfy. Note the ceiling as well as the floor: a single concept is below comparative mode, and five or more switches the guide to a card sort, which does not produce these sections either. Two to four is the band that gets you the comparison.
In the head-to-head read, each line under What drove the choice carries a count taken from the interviews. A reason people gave whatever they chose reads like 4 of 12 participants. A reason tied to one option is counted only among the people who chose it, and reads like 5 of the participants who picked followed by the option. A reason with no count shows none.
If you ran a name, value proposition, or messaging test rather than a standard one, the report is titled by that sub-type, for example Value proposition test v2. Version numbers count within a sub-type, so your first name test is v1 even if standard concept tests already ran in that study. The report structure is the same; what changes is that the concepts being compared are options for one anchor concept rather than peer ideas.
Comparative reports count each participant's favourite out of everyone in the report, in the decision table's Favourite column, for example 18 of 24. On a name, value proposition or messaging test, and on a test run from an uploaded interview guide, the cell also says how many picks were reluctant, as in 18 of 24 (2 reluctant), as long as every pick on the report recorded it. A standard comparative test never shows a reluctant count. A reluctant pick means the participant said none of the options appealed, then named that one as their choice if they had to pick. Someone who names no option is not counted for any option. Reports made before this count existed show the older count, out of the picks alone.
Good, and this is the one label in any Candor report most likely to be read backwards. Severity describes how hard the objection is to answer, not how loudly your synthetic participants raised it. Low severity in green means the concept already has a strong answer to that objection, so it's addressable with better wording or a feature you could add. High severity in red means the concept has no good answer, which makes the objection structural rather than a messaging problem. A concept with several High-severity objections needs rethinking, not rewriting.
Drop means the idea didn't land. Rewrite-and-retest means you can't tell yet, because the description got in the way. It's the only verdict coloured blue rather than green, yellow, or red, precisely because it isn't a judgement on the concept: it's a judgement on how the concept was written. Treat it as a signal that the test is unfinished. Rewrite the concept statement and run it again rather than reading it as a soft no.
Because the run wasn't in comparative mode, and there are two ways to miss it. One concept is below the threshold. Five or more switches the guide to a card sort, which ranks concepts but doesn't produce the head-to-head read, the What would have to be true to adopt section, or the B2B approval requirements. Comparative mode is two to four concepts, so that is the band to aim for when the point of the study is choosing between directions. Outside it you still get verdicts, comprehension, objections, and next steps per concept.
Start with a free project, or walk through Candor with us first.