Mutterbox

Sample report

What a report looks like

Below is a real report generated by Mutterbox from a demo venue. It is the same file a client receives — nothing here was prepared specially for the website.

The sample report is not shared at the moment. Create an account — the demo venues have a full review history and you can run an analysis right away.

Contents

Six parts of the report

  1. A verdict on page one

    One sentence on where the venue stands and how many problems are growing. The rest of the report expands on it.

  2. Review statistics

    Review count, average rating, star and sentiment distribution — with an explicit note on how many reviews had no text.

  3. Strengths

    Areas praised more often than criticised, with mention counts. Useful for communication, not only for repairs.

  4. Problem cards

    Recurring complaints with occurrence counts, first and last date, trend, confidence and verified quotes.

  5. Recommendations across three horizons

    Within 7 days, within 30 days, beyond 30 days. Each with priority, cost, time required and expected effect.

  6. Methodology

    Where the data comes from, how the trend is calculated and what the report does not cover. Without this part, the rest cannot be checked.

Anatomy

What exactly sits in a problem card

Condition of the pool infrastructure

Critical rising Confidence: high

94 of 211 reviews with complaints · latest 18.07.2026 · oldest 14.04.2012

Despite generally positive ratings, visitors repeatedly raise concerns about closed slides and low water temperature in the pools.

„the slides were closed and the ticket cost the same as always." Google · 18.03.2026 · review #3187
  • The occurrence count comes from the database

    It is the result of an SQL query over classified signals, not a number supplied by the model. The model does not see this number while writing the description.

  • The quote is cut from the original

    We store the character position inside a specific review. What reaches the report is a fragment cut from the original text — typos of the author included.

  • Trend measures share, not sum

    In a growing venue everything grows, complaints included. So we compare the share of complaints among reviews from each period.

  • Confidence says how much to trust it

    It depends on the number of source reviews, their spread over time, and whether more than one source confirms the problem.

Blog

More on the method

Be first in when we launch

Join the waitlist. We'll write once, on launch day, with a link to create your account.