Back to Analyse Documentation

How to use Sentimento

Sentimento turns a spreadsheet of customer reviews into clear, actionable insight. Upload a CSV or XLSX file and we automatically classify the sentiment of every review, surface the themes people keep mentioning, and write you an executive summary. This guide walks you through preparing your file and reading your results — no account required to read it.

What Sentimento gives you

Sentiment per review

Every review is classified as Positive, Negative, Mixed or Neutral based on its overall tone.

Themes & topics

We detect the topics people mention — quality, delivery, price, customer service, packaging and more — and count how often each one comes up.

Executive summary

A written overview of what's working, what needs attention and the key actions to take, generated from your full dataset.

Exportable results

Download the complete analysis — sentiment, themes and summary — as a formatted XLSX workbook to share with your team.

Step 1 — Prepare your file

Your file needs just one column of review text, with one review per row. The first row is the header. Name that header any of the accepted column names below and Sentimento will find it automatically — anything else in the file is ignored.

The top of a typical reviews file. Column A is headed review and holds one review per row — that's all Sentimento needs. Extra columns like date or rating are fine; they're simply ignored.

Accepted header names

Case-insensitive. Use any one of these:

reviewreviewstextcommentfeedbackdescriptionreview_textreview textreviewtext

File format

  • .csv (UTF-8 recommended) or .xlsx
  • First row must be the header
  • One review per row in the review column
  • Empty rows are skipped automatically

Size limits

  • The 14-day free trial covers up to 100 reviews per analysis
  • Paid plans raise the per-analysis limit
  • Your current limit is shown on the upload screen
See the raw CSV text
review,date,rating
"The product quality is excellent, but shipping took way too long.",2026-01-04,3
"Excellent service and the packaging was beautiful. Will buy again!",2026-01-05,5
"Not impressed — poor value for money and support never replied.",2026-01-05,2
"Works as expected. Nothing special but it does the job.",2026-01-06,4

Tip: wrap reviews that contain commas in double quotes so each review stays in a single cell. Most exports from Excel, Google Sheets, Trustpilot, Shopify, Zendesk and survey tools already do this for you.

Step 2 — Run your analysis

  1. 1

    Create a project

    Projects keep related analyses together — for example one project per product, brand or month. Create one from the Projects tab before your first upload.

  2. 2

    Open Analyse and pick your project

    Go to Analyse, choose the project, and give the analysis a memorable name (e.g. “January Trustpilot reviews”).

  3. 3

    Upload your file

    Drag your CSV or XLSX onto the upload zone, or click to browse. Sentimento detects the review column and counts how many reviews it found before you start.

  4. 4

    Start the analysis

    Click Start Analysis. Reviews are processed in batches and a live progress bar shows how far along it is — you can leave the page and come back.

Step 3 — Read your results

When the status turns completed, open the analysis to explore it. You'll see:

  • Sentiment breakdown — how many reviews fall into Positive, Negative, Mixed and Neutral.
  • Theme map — the topics customers mention most, and the sentiment attached to each.
  • Executive summary — what's working, what needs attention, and recommended actions in plain language.
  • Per-review detail — each review with its sentiment and the exact phrases that drove the classification.
  • Export — download everything as an XLSX workbook to share or archive.

If processing is interrupted, any reviews already analysed are saved and the analysis is marked partial so you never lose work.

Frequently asked questions

Why does it say “No reviews found”?

Your file's header row probably doesn't use one of the accepted column names. Rename the column holding your review text to review (or any name from the list above) and upload again.

Do I need a paid subscription to read this guide?

No. This documentation is public and available to everyone. You'll need an account and an active trial or plan to actually run an analysis.

What languages are supported?

Sentiment and theme detection work best on English reviews, but the underlying model handles many languages. For non-English data, spot-check a few results first.

Can I include other columns like date or rating?

Yes. Extra columns are ignored, so you can upload an export straight from your review platform without trimming it down.

What if my file has more reviews than my plan allows?

Split it into smaller files within your per-analysis limit, or upgrade your plan to raise the limit. Your current limit is always shown on the Analyse screen.

Ready to analyse your reviews?

Head back to the upload screen and drop in your file.