Reading a synthesis report

Where reports come from

You don’t launch synthesis yourself. When every session in an auto-interview run reaches a terminal state, Candor starts synthesis automatically and the progress bar on the run page shifts to the synthesis phase. When it finishes, a banner appears with a View report link.

One run produces one report. Running the same guide twice gives you two reports, and both are kept.

The shared spine

All four interview types lay out the same way, in four sections. Only the contents of Findings really change.

• What Matters Most. The headline read on the run, and the only section expanded when the report opens. See below.
• Findings. The substance, and the part that differs most by interview type, in what it holds and in how it is ordered. Your type’s article covers both.
• Next Steps. The sequenced action plan. In discovery and validation reports each move links back to the findings it came from. Price testing calls this section What’s Next and frames each move as a test to run rather than a finding to revisit.
• Evidence appendix. Labelled “dig deeper into the evidence if interested.” This is where the supporting detail and the full transcripts live when you want to check the work rather than act on it.

Everything except What Matters Most starts collapsed. That is deliberate: the report is meant to be skimmed to a decision first and audited second. Cross-reference links will open a collapsed section for you when you follow one.

Start with the learnings block

Inside What Matters Most, alongside the headline and the sample details, is a learnings block. It is the part you should always read, and it’s worth knowing that it appears in a slightly different form depending on the interview type.

• Problem validation, concept testing, and price testing. The block is titled Top Learnings & Recommended Next Steps and pairs each learning with the step it implies, so you see what was found and what to do about it side by side.
• Problem discovery. The block is titled Key Learnings and deliberately carries no per-learning action. Discovery is for understanding, so the actions live in Findings and Next Steps instead. Every report type, discovery included, also calls out the single highest-priority move as Do this first.

In problem discovery and problem validation reports, each learning also carries clickable links down to the findings it was drawn from. Use those rather than scrolling: they take you straight to the evidence for that specific claim.

Confidence levels

Where a report shows a confidence level, it runs High, Medium-High, Medium, or Low. The colour cools as confidence drops, so you can see a weak area without reading the labels. Not every report attaches one to every finding: concept testing leads with a verdict per concept instead.

Confidence describes how much weight the evidence can carry, not how important the finding is. A Medium finding can be the most consequential thing in the report.

When something pulled a level down, an info icon appears next to it. Open it and you get the specific caps that fired, each with its reason. That popover is the useful part: it tells you what to change in the next run to raise confidence, rather than leaving you to guess.

Verdicts

Some report types return an explicit verdict as a coloured pill. Each type uses its own vocabulary, so the words differ by report.

The words differ by type, so your type’s article lists its own set. The colours do not, because every report draws them from the same palette: green means go, amber means change something, red means stop, grey means the run couldn’t tell you, and blue means the verdict is about how something was described rather than about the thing itself.

When the banner says the run was thin

An amber banner at the top of a report means the run was partial or thin in a way that affects how far you should push the conclusions. It lists each issue on its own line, with the personas involved where that applies. The usual causes:

• Too few personas. Two or fewer sessions completed, so nothing can be called a pattern.
• Thin transcripts. One or more sessions ended before eight exchanges, which is too short to reach the substance.
• Goals were inferred. The study had no explicit learning goals, so Candor worked out what you were probably asking. Flagged for transparency, not as a failure.
• Sessions excluded. Some sessions were filtered out before synthesis ran, usually because a quality check rejected them.

The report is still worth reading when this banner appears. Treat it as a narrower base, and if the decision is expensive, run again with more personas before committing.

What changes by interview type

Everything above holds for any report you open. What each section actually contains does not, and neither does the verdict vocabulary or what lands in the appendix. Each type has its own article covering its sections and the judgement calls specific to reading it.

• Reading a problem discovery report. Key Learnings, What Surprised Us, findings ranked by reach, and the unprompted signal.
• Reading a problem validation report. The five verdicts, tested hypotheses, and the problems that surfaced on their own.
• Reading a concept testing report. The decision table, the misread and name alerts, and why objection severity reads backwards.
• Reading a price testing report. Recommended tiers, the willingness-to-pay chart, anchors, and stated versus revealed.

Exporting

The export menu in the report header offers three scopes.

• Report. The written report on its own, as Markdown or PDF.
• Transcripts. The raw interviews, as CSV, Markdown, or PDF. CSV always comes back as one file with every session in it. Markdown and PDF come back as a zip with one file per persona.
• Everything. A zip holding the report plus every transcript, as Markdown or PDF.

Reach for Markdown or CSV when you want to do your own analysis elsewhere. Neither truncates, so you can hand the whole run to another tool. PDF is the one to send to someone who just needs to read it.

When synthesis fails

Synthesis runs as a chain of steps. If one fails, the report page shows which step it was and gives you a Retry from failed step button, so a retry picks up where it stopped instead of starting over. Most failures are transient and a single retry clears them.

Where to go next

• Running an auto-interview
• Choosing an interview type
• What “evidence-grounded” means in Candor