Open-ended survey analysis alongside reviews

Why combine them

Open-ended survey analysis on its own is shaped by the questions you asked and the customers who answered. Reviews are unprompted but come from customers motivated to write publicly. Put together, they cover each other's blind spots.

One taxonomy

Use the same themes and sub-themes for survey comments and reviews. Without that, "checkout" in one and "billing" in the other look like different topics.

Use the scores

• Split survey comments by score: detractors, passives, promoters (or low, middle, high CSAT).
• For each topic, compare how often detractors and promoters mention it.
• Topics far more common among detractors are likely drivers of dissatisfaction.

Compare with reviews

• Agreement: the same topic rising in survey comments and reviews is strong evidence.
• Surveys only: a topic customers mention when asked but rarely volunteer; real but maybe less urgent.
• Reviews only: a topic customers feel strongly enough to publish; check whether your survey asks about it.

An example

A hotel group's survey comments from detractors mention breakfast crowding; reviews mention it too, at the same properties, at weekends. Both sources agree, so breakfast capacity at weekends becomes a priority. Reviews also mention noise from events, which the survey never asked about; a question is added.

Common mistakes

• Separate categories for surveys and reviews.
• Reading comments without their scores.
• Combining sources into one number.
• Ignoring topics that appear in only one source.

Common questions

Checklist

• One taxonomy across sources.
• Comments split by score.
• Sources reported separately.
• Agreement and gaps noted.

In short

Classify survey comments and reviews with one taxonomy, compare detractors with promoters, and treat agreement across sources as strong evidence and gaps as questions.

A step-by-step method

Export three months of survey responses with their scores and comments, and three months of reviews for the same locations. Classify both with your taxonomy at the mention level. For each topic, calculate how often it appears among detractors, among promoters and in negative review mentions. List the topics where detractors and negative reviews agree; those are your priorities. List the topics that appear in reviews but never in surveys; consider adding a question.

Keep the two sources in separate columns in every table, so the comparison stays visible and a change in survey response rates is not mistaken for a change in experience.

Things to watch

Survey timing matters: a survey sent minutes after a visit captures different things from a review written that evening. Survey questions also prime answers; a question about staff produces staff comments. Note these differences when comparing, rather than treating the two sources as identical.

Reporting the combined view

Present a simple table per topic: share of detractor comments mentioning it, share of promoter comments mentioning it, negative share in reviews, and the change in each since last period. Add two quotes from each source. Leadership can then see at a glance which topics both sources point to, which only one does, and how strongly.

Doing this with Falcon

Falcon builds the theme list from your own reviews, then labels every review against it with a multilingual model, scoring sentiment for each mention and keeping the exact quote behind each label. Texts too short to judge are marked unclear rather than guessed, and low-confidence labels are flagged for a person to check. Negative share by topic, location and month then shows what is rising where, with small samples flagged.