Reading the results

Once responses come in, Prism turns them into something you can actually read and act on.

The numbers and the "why"

Every question gets a chart — a distribution for a scale, counts for a choice, a word view for open text, and for a MaxDiff set, the items ordered by how often each was picked best minus how often it was picked worst. Alongside them, when you analyze the study, Prism reads the responses and pulls out themes: the patterns that show up across people. Each theme carries the real quotes behind it, so you're never asked to take a summary on faith.

You can analyze again as more responses arrive — Prism picks up the new answers and refines what it found, rather than starting over.

Where people stopped

Above the charts is the shape of the whole pass: how many opened it, how many started, how many finished, and how long the median person took. Biggest drop-off names the question that lost you the most people.

Each question card carries the same story locally — how many answered it, how many saw it, and a mark when the drop is steep. A question that bleeds a third of your sample is usually a question that's too long, too personal, or asked too early, and the number tells you which one to look at before you field the next wave.

A question answered by fewer than thirty people is marked Low base. Read those as a direction, not a finding.

Weighting

If the people who answered aren't the people you care about, weighting rakes the sample to shares you set — age, role, segment — so the headline numbers speak for the market rather than for whoever clicked. Prism shows the effective sample size alongside the real one, so you can see what the weighting cost you. Rankings, MaxDiff and free text stay unweighted.

Cutting the data

Crosstabscompare groups

Split any question by another — satisfaction by customer type, for instance — and see the groups side by side. Prism tells you when a difference is large enough to trust, and stays honest when a group is too small to say either way.

Segmentsdiscover groups

Ask Prism to find the natural groupings hidden in the answers. It surfaces the segments that emerge, names them from what actually distinguishes them, and shows you how they differ.

The report

Turn a study into a shareable report: a written summary of the key findings and recommendations, above the charts that back them. It's print-ready, so you can hand it to someone who'll never open Prism.

You can also export the raw responses, a topline of every number, or a crosstab as a spreadsheet — and interview transcripts as a document — whenever you'd rather work with the data yourself.

Keeping what you learn

A finding is only worth something if it sticks.

Save it to Ledger. Promote a theme and it becomes a belief your team remembers, with the quotes that support it and a link back to the study.
Track it over time. Set a study to repeat on a schedule, and each round posts its score to a running trendline — brand sentiment, NPS, whatever you're watching — so you see the number move, not just a one-off snapshot.
Ask the next question. Every analysis suggests what to explore next, and you can draft the follow-up study from it in a click.

4 min read