Sopact is a technology based social enterprise committed to helping organizations measure impact by directly involving their stakeholders.
Copyright 2015-2026 © sopact. All rights reserved.
A training feedback survey captures reaction at session end. What it measures, what it cannot, and how to design one that feeds real evaluation — with examples.
A training feedback survey is a short questionnaire given at the end of a session that captures how participants reacted to the training - its relevance, content, facilitator, and pace - so you can judge whether it landed and improve the next run. It is Kirkpatrick Level 1 (Reaction), the first and most common measure of training evaluation.
Most feedback surveys are treated as a self-contained smile sheet: collect the ratings, average the stars, move on. Designed well, the same survey is the opening move of a connected evaluation. With Sopact, each response is tied to a persistent participant ID and the questions are written to map forward, so the reaction you collect today connects to the learning, behavior, and results you measure at Levels 2-4 on the same record.
Used by: L&D and training teams, workforce and skills programs, HR and people analytics, and anyone who has to report not just that a session ran but whether it was worth running again.
A standalone smile sheet ends the moment the ratings are averaged. Every send is a fresh, anonymous batch, so the 4.3-out-of-5 you collect today can never be joined to whether the same person actually learned or applied anything. That is why so much training evaluation is a stack of Level 1 averages and nothing else - the survey was designed to collect a score, not to connect a record.
The fix is a design decision, not a longer survey. In Sopact, a training feedback survey is built as Level 1 of a connected chain: every response carries the participant's persistent ID, and every question is chosen because it maps forward to something you will measure later. Relevance today predicts application at Level 3. A low reaction score flags who may disengage before the Level 2 pre/post even runs. Because the ID persists, the reaction survey, the learning assessment, and the 90-day behavior follow-up all land on one record instead of in three disconnected tools. For the wider evaluation that this feeds, see training evaluation.
Keep it short and keep it forward-looking. The core reaction set covers five closed questions plus one open question, and each one is chosen for what it will connect to later, not just for a number on a slide.
Relevance to the job - How relevant was this training to your day-to-day work? This is the single best early predictor of whether the skill gets applied at Level 3, so it earns its place. Content quality - How clear and useful was the material? Facilitator - How effective was the facilitator at explaining the content? Pace - Was the pace too slow, about right, or too fast? Pace is the quiet killer of a good session and the theme most likely to hide behind a mediocre score. Likelihood to recommend - How likely are you to recommend this training to a colleague? - a single-number confidence read.
Then the one question that does the heavy lifting: an open-ended most / least useful prompt - What was the most useful part of this training, and what was the least? The star ratings tell you the score; this open question tells you the reason behind it, and it is the raw material Sopact codes into themes and satisfaction drivers. Read the comments, not just the stars - the theme behind a low rating (pace, relevance, a confusing module) is what you actually act on. For the fuller bank, see employee training survey questions and the wider training feedback practice.
The reason to design the survey this way is that each Level 1 answer has a job further down the chain. Relevance and the open most/least-useful comment map to Level 2 (Learning): the topics rated least useful are the ones to check for a weak pre-to-post gain. The reaction score and drop-off flags map to Level 3 (Behavior): a low-reaction participant is the one whose 90-day application you watch, and the barriers named in the open question often reappear as transfer barriers at work. The participant ID itself maps to Level 4 (Results): because the same record carries reaction, learning, and behavior, a board-ready result can be traced back through every level. Design the questions once, tied to one ID, and the smile sheet becomes the first link in a chain instead of a dead end. The full four-level view is on training feedback.
Analysis has two halves that a spreadsheet keeps apart and a connected survey keeps together. The closed questions give you the numbers - a reaction score per participant, distributions on relevance, facilitator, and pace. The open most/least-useful question gives you the reasons, and this is where most teams stall: they read a handful of comments and guess at the rest. Sopact codes every open-ended response into consistent themes on the same participant ID, ranks the top drivers of low satisfaction, and pulls a representative quote for each - so the score and the reason sit on one record. The mechanics of coding open text are covered in how to analyze survey data.
Because every answer is tied to the participant ID, the analysis does not stop at this survey. A reaction that predicts drop-off is flagged for the Level 2 follow-up; a barrier named in the open comment is carried into the Level 3 behavior check. That is the difference between analyzing a smile sheet and running Level 1 of a connected evaluation. The wider discipline this feeds is impact measurement & management.
A short walkthrough of collecting and analyzing training feedback on one connected record.
Design the reaction set, collect it on one link tied to each participant ID, code the open comments into themes and drivers, then connect the reaction to the record so Levels 2-4 can build on it. The animation below shows a smile sheet turning into the first link of a connected evaluation; the four prompts under it are the ones you paste into the Sopact Assistant.
1 - Design and analyze the reaction survey. Build the Level 1 instrument and code the feedback into a score plus drivers on each participant ID. The walkthrough is in how to design a training feedback survey.
Academy walkthrough → Design a training feedback survey
From this training feedback for [COHORT], code the open-ended most/least-useful responses into themes (content, facilitator, relevance, pace), give a reaction score per participant on their persistent ID, rank the top two drivers of low satisfaction with a representative quote each, and flag reactions that predict drop-off.
2 - Code the open comments. Turn the free-text answers into consistent themes rather than reading a handful and guessing. The walkthrough is in analyze open-ended survey responses.
Academy walkthrough → Analyze open-ended survey responses
Code the open-ended responses in this training feedback into a consistent set of themes, report the frequency of each theme, and attach a representative verbatim quote to each - keeping every response tied to its participant ID so the themes connect to later levels.
3 - Rank the satisfaction drivers. Find which themes actually move the reaction score, not just which appear most often. The walkthrough is in analyze sentiment and satisfaction drivers.
Academy walkthrough → Analyze sentiment and satisfaction drivers
From this training feedback, score sentiment per participant and identify the top drivers of low and high satisfaction, ranking them by how strongly each is associated with the overall reaction score, with a representative quote for each driver.
4 - Connect reaction to the four levels. Set the feedback survey up as Level 1 of the full evaluation so it links forward to learning, behavior, and results on one ID. The walkthrough is in apply the Kirkpatrick model to a survey.
Academy walkthrough → Apply the Kirkpatrick model to a survey
Set up this training feedback survey as Level 1 of a four-level Kirkpatrick evaluation on one persistent participant ID: capture the reaction, and design the flow so the same ID carries into the Level 2 pre/post assessment, the Level 3 behavior follow-up, and the Level 4 results pull, keeping all four levels on one record.
The sections above are the argument; the Academy articles are the practice - each written to run on your own cohort data, starting with the survey design itself.
A training feedback survey is a short questionnaire given at the end of a session that captures how participants reacted to the training - relevance, content, facilitator, and pace. It is Kirkpatrick Level 1 (Reaction). Designed well, it is not a standalone smile sheet: in Sopact each response is tied to a persistent participant ID so the reaction connects forward to learning, behavior, and results on the same record.
Include five closed questions - relevance to the job, content quality, facilitator effectiveness, pace, and likelihood to recommend - plus one open-ended most/least-useful question. The open question tells you the reason behind the score. In Sopact each question is chosen because it maps forward: relevance predicts Level 3 application, and a low reaction flags who to watch at Level 2, all tied to one participant ID.
A smile sheet ends when the ratings are averaged; every send is a fresh anonymous batch that can never be joined to what the same person learned or applied. A connected training feedback survey, built in Sopact, ties every response to a persistent participant ID and chooses questions that map forward to Levels 2-4, so the reaction is the first link in a connected evaluation rather than a dead end.
Analyze the closed questions for a reaction score and distributions, and the open most/least-useful question for the reasons behind the score. Sopact codes every open-ended response into consistent themes on the same participant ID, ranks the top drivers of low satisfaction, and attaches a representative quote to each - so the number and the reason sit on one record and feed the later Kirkpatrick levels.
Anonymity breaks the chain. If the reaction survey is anonymous, you can never connect it to the same participant's Level 2 learning gain or Level 3 on-the-job behavior. Sopact ties every response to a persistent participant ID assigned at enrollment, which keeps the survey confidential in reporting while still letting reaction, learning, behavior, and results live on one record for a defensible evaluation.
The feedback survey is Kirkpatrick Level 1 (Reaction). Its value multiplies when it is designed as the opening move of the full four-level model: relevance and the open comment map to Level 2 (Learning), the reaction score and drop-off flags map to Level 3 (Behavior), and the shared participant ID lets a Level 4 (Results) number be traced back through every level. Sopact runs all four on one record.
The single most useful open question is a most/least-useful prompt: What was the most useful part of this training, and what was the least? It surfaces the reason behind the rating and the barriers a participant may hit later. In Sopact this free text is coded into themes and satisfaction drivers on the participant ID, so it becomes structured evidence rather than comments no one reads.