When audience generation finishes, Candor drops you into a review screen with a lot on it. This article walks through every section, what each badge and field means, and what to look for before your synthetic participants get built.
The header shows your project name, the audience type (B2C or B2B), and the regions you pinned during setup. These are quick reference; the actual decisions happen below.
Normally your participants start building the moment the audience is finished, and you never see this screen sitting idle. Occasionally that hand-off doesn’t happen. When it doesn’t, a short message appears at the top of the screen saying so and saying what to do about it. If you don’t see that message, nothing went wrong.
Most of the time it asks you to pick your groups and press Build participants yourself, and that is all it takes. Your audience is finished either way and none of it is lost. Some versions point you somewhere else instead, because pressing the button would run into the same wall: a payment that hasn’t gone through, which links you to your billing settings, and a workspace that’s been paused, which asks you to contact us because only we can lift that. Once the payment or the pause is sorted, the message changes to say so and invites you to go ahead.
There is one more version, and it is the one that says the least on purpose. To tell you whether a payment or a pause still stands, we have to look at your workspace as it is right now, and once in a while that lookup doesn’t come back. When that happens the message says we couldn’t check, rather than guessing. It won’t tell you the problem is still there and it won’t tell you it has been sorted, because at that moment we don’t know either. It still tells you the true part, that this study has no participants yet, and it asks you to reload in a moment or contact us if it keeps happening. It deliberately stops short of telling you to press Build participants, since that is exactly the thing we couldn’t confirm would work.
Whether that screen has a picker at all is decided by your audience, not by the message. Building participants needs at least two groups, so if yours came back with fewer there’s no selection that gets past it: the screen shows the groups it did find and a way to reach us, instead of a picker and a button. If a message tells you to pick your groups and there’s no picker, that’s why.
When there is a picker, it has the same floor. Untick your way below two groups and Build participants greys out, with a note beside it saying how many you need. Tick one back on and it comes alive again.
What it says depends on what you’re allowed to do. If you can’t open billing settings, it names who to ask instead of showing you a link that goes nowhere. If your role can view a study but not run one, it asks you to pass it to someone who can, rather than pointing you at a button that would turn you away.
Further down the page, under the signal list. Most signals are lifted from one page and show that page as their source. Some are pulled together from two or three sources at once, and those list each source rather than one. Pulling findings together is where a claim is most likely to bend, so Candor re-reads each of those against the wording of the sources it cites.
On an ordinary run the page shows nothing about this check: no count of claims checked and no count of claims that didn’t hold up. Not holding up means the claim bends what its sources say. A number that measured one thing described as another. A study’s country or setting dropped. Two figures with different bases merged into one range.
A line above the list appears only when the check didn’t run, or found no synthesized signal it could trace back to its sources. It says those signals are unverified rather than confirmed, without marking which ones. Read it when it appears, because a run that checked nothing looks exactly like a run where everything held up. Treat it as a reason to weight the synthesized claims lightly, or to add stronger evidence and regenerate. Nothing is removed either way.
The check isn’t the end of it. During the run, every claim that doesn’t hold up is sent back once to be rewritten against its own sources, and the rewrite is kept only if re-running the check rates it better. A claim Candor can’t improve is kept exactly as first written, and nothing is deleted either way.
Each group is a meaningful slice of your audience, defined by a primary differentiating dimension. That’s the specific behavioral or psychographic criterion that explains why members of this group make different choices from others. It’s not a demographic split. Two same-aged buyers in the same city can sit in different groups because they evaluate decisions differently.
On each card you’ll see:
This is where the depth lives. Click the supporting-signals footer and the group expands to show every signal that fed it. For each signal you get:
If a group’s signals are mostly assumed, that’s the kind of group to scrutinise before including. The participants you generate from it will be necessarily speculative. There is no per-signal confidence figure on this row: it was dropped because a percentage on every one of several hundred rows is noise, and the provenance label already tells you how solid the claim is.
At the foot of the page, collapsed until you open it. Two things sit here. Key dimensions modeled is the set of behavioural dimensions Candor picked to tell your groups apart. Personality and bias calibration holds OCEAN priorities (which Big Five traits matter most for this audience, ranked high, medium or low) and primary biases (the cognitive biases this audience shows most). These are the levers that’ll show up in the participants.
You don’t pick. Candor generates your audience and its participants in one run, and every group with enough evidence behind it becomes participants automatically. Groups Candor flagged as low-confidence are left out, and the population shares of the rest renormalise so they still add up.
One exception keeps a run from being wasted: participants need at least two groups to work with, so if fewer than two clear the evidence bar, Candor keeps them all rather than stopping. A thinner audience is more useful than none, and the evidence depth on each group card tells you which ones to read with caution.
Participants you don’t generate can’t be added later inside the same project, which is why the description you write at setup does the real work here. It decides which groups exist in the first place.
One gate to know about: if a recent payment for your workspace didn’t go through, new generation runs are paused until billing is settled. Update your card in Settings → Billing and try again. Everything you’ve already generated stays available to view.
Two things worth checking, and both point back to your setup inputs rather than to anything on this screen:
Start with a free project, or walk through Candor with us first.