Collect event feedback — after the event

Sending a survey right at the event end is intuitive — but rarely optimal. In the first hours residual fatigue or travel stress dominate, answers become hasty and superficial. After 24 to 48 hours the impression has settled, the context is still fresh. From day 5 onwards the memory of individual sessions noticeably fades.

For hybrid or online events the timing is even tighter. Here you lose attention twice as fast — the context switch back to daily work happens immediately. A first short survey within two hours with a single question ("How did you like the event?") secures the initial signal. The detailed survey follows 24 hours later.

Automation is mandatory: manual sending the next day gets forgotten when the event team is busy with wrap-up. A delayed webhook after event end or a cron trigger at a fixed time are the cleaner solutions. Do not forget the reminder after three days for non-responders — it typically lifts the response rate by another 15 to 25 percent.

Pure scale surveys ("How did you like session X, 1-5?") deliver measurable data but no understanding. Pure free-text surveys deliver understanding but cannot be aggregated. Mixing both in a fixed ratio works best in practice — and still keeps the survey short.

A proven pattern for event feedback: per session two closed questions (content rating, presentation rating) plus one open question ("What would you have wished for additionally?"). At the end of the event three overall questions: NPS, top-3 highlights as multi-select, one open improvement question. With conditional logic you can also query only sessions the participant actually attended — e.g. via hidden field from the registration tool.

The order matters. Start with the closed questions — they are quick to answer and build momentum. Open questions come at the end because they are cognitively more demanding. Anyone who clicked the scale at 4 and 5 is more willing to write two sentences afterwards — anyone forced to start with free text drops off more often.

A common reflex after an event is the question "Which session was best?". Simply asking this openly produces a recency bias — the last session gets disproportionate mention. Structured comparison needs a different setup.

The cleanest method is to rate each session individually and aggregate afterwards. Per session: two or three uniform questions (content, presentation, practical relevance on a 1-5 scale). At the end an additional multi-select "Which three sessions did you take the most away from?" as secondary validation. Both data sources combined give a robust ranking.

Watch out for an anti-bias: do not cut off too early. If you only show top ratings in a report, you only see half-truths. A session averaging 4.5 with high variance (some love it, some hate it) is a different phenomenon than a session averaging 4.3 with tight spread. Both values belong in the report — the mean rarely suffices alone.

Event NPS ("How likely are you to recommend this event?") is the only metric you can compare across multiple events. A pure satisfaction scale varies too much because expectation levels differ per event. NPS normalises that because the recommendation question has a clear external reference.

The evaluation follows the standard: promoters (9-10) minus detractors (0-6) as percentage points. It is important to display not only as a score but also as a distribution — an NPS of 30 from "many 9s, some 6s" is different from "many 10s, many 5s". The latter is a polarisation signal and needs different actions.

Always pair NPS with an open follow-up: for promoters "What was the highlight for you?", for detractors "What was missing?". These qualitative answers are the actual goldmine — they tell you why the numbers are what they are. An AI summary per segment then gives the quick overview without anyone having to read 200 free-text answers.

An event feedback evaluation that only ends up as a PDF in a folder was a waste of time for the participants. They invested in giving answers — you owe them at least visibility into what came of it. The improvement of the next edition is the honest currency.

Three concrete actions per evaluation are the minimum: a short mail to all participants ("This is what we learned, this is what we change"), an internal list with three measurable improvements for the next event, a letter to low-rated speakers with concrete feedback (without respondent names but with tendency). The latter is uncomfortable but necessary — speakers can only improve when honestly mirrored.

Measure across multiple events whether your actions work. If at the next event the same issue is criticised again, the measure was either insufficient or the diagnosis was wrong. An activity log per event with the actions taken creates continuity, especially when the event team rotates — otherwise every new edition starts at zero and the learning curve flattens.