What this prompt does
This prompt builds a customer feedback analysis framework that turns scattered comments across channels into structured, de-duplicated signal. You provide [product_name], the [feedback_channels] you collect from, your monthly [feedback_volume], and your [customer_segments]. ChatGPT then designs a categorization taxonomy, a sentiment scorer, a feature-demand aggregator, a trend analyzer, a competitive-mention tracker, an executive summary template, and a response-suggestion system.
The variables tune the analysis to your reality. [category_count] sets how many top-level buckets feedback gets sorted into, [weighting_criteria] decides how feature requests are weighted (by ARR, segment, health score), and [threshold_shift] defines how big a sentiment swing must be before it gets flagged. [comparison_period] anchors trend detection, [competitor_names] tells the tracker who to watch for, and [report_frequency] and [escalation_team] shape the reporting and routing. The aggregation step is the core value: it groups requests worded differently into one real feature need and counts unique requesters rather than raw mentions.
When to use it
- Feedback is pouring in from several channels and the signal is buried until something organizes it.
- You want feature requests de-duplicated so "dark mode," "too bright," and "eye strain" collapse into one need.
- You need a consistent categorization scheme so different people tag feedback the same way.
- You want early warning when a category's sentiment shifts before the issue becomes widespread.
- You need a recurring executive summary of themes, NPS trend, and urgent issues for product reviews.
- You want competitive mentions surfaced so you know which rival capabilities customers compare you to.
Example output
You get a structured analysis framework. It includes a categorization taxonomy with [category_count] top-level categories plus subcategories and example phrases, a sentiment-scoring spec on a -5 to +5 scale that tags the product area and pulls verbatim quotes, a feature-demand aggregation method that groups synonyms and counts unique requesters weighted by [weighting_criteria], a trend-analysis approach comparing against [comparison_period] and flagging shifts beyond [threshold_shift] points, a competitive tracker for [competitor_names], a [report_frequency] executive summary template, and a response-suggestion system routing urgent items to [escalation_team].
Pro tips
- List every real source in
[feedback_channels]so the taxonomy and routing account for the formats each channel produces. - Keep
[category_count]manageable — too many top-level buckets makes tagging inconsistent, which defeats the de-duplication goal. - Choose
[weighting_criteria]you can actually pull; weighting by ARR is powerful only if you can join feedback to account data. - Set
[threshold_shift]to a level that flags real emerging issues without drowning you in noise from normal monthly variance. - The framework defines how to analyze feedback, but running it at scale needs the model fed actual feedback batches, so plan that step.
- Keep
[competitor_names]current — an outdated competitor list means you miss mentions of whoever is actually pulling customers away.
