Problems
You have reviews in a spreadsheet and stopped filling it in
Almost every venue we speak to has a spreadsheet of reviews somewhere. In almost all of them the last entry is six months old.
Why the spreadsheet always dies
It requires three actions per review: copying it, assigning a category and judging the sentiment. The last two are subjective, so the data stops being comparable the moment a different person fills it in. The first is simply dull. The result is always the same: the columns are there and the rows stopped months ago.
The category depends on the mood
The same review lands once under "service" and once under "waiting time", depending on who typed it and on which day. After a year of that, the shares mean nothing.
No quote, just a summary
What goes into the cell is a summary, because the full review does not fit. The summary is made by the same person who later draws conclusions — from their own summary, not from the review.
The history never closes
The spreadsheet starts on the day it was created. Reviews from three years ago stay untouched, and those are exactly the ones showing whether last year's refurbishment changed anything.
Step by step
What survives of the spreadsheet and what does not
-
1
Copying disappears
Reviews no longer go into the spreadsheet by hand. The CSV export delivers them with date, platform, rating and full source text — ready to paste.
-
2
Subjective categories disappear
Classification is consistent, so percentage shares start to mean something. With manual entry the same review landed in different places depending on the person.
-
3
Your notes stay
Columns with decisions, deadlines and task assignments stay where they were. That is what a spreadsheet is genuinely good at and nothing replaces it.
-
4
The history closes
The analysis pulls the full review history, including from before the spreadsheet existed. That is how you see whether last year's refurbishment changed anything.
In practice
Why a spreadsheet cannot show a trend
It starts on the day it was created and ends on the day somebody stopped filling it in. Between those dates the data is only as good as the diligence of whoever typed it. The analysis reaches back to the first review and classifies all of them the same way — only then does comparing periods mean anything.
from the first
review, not from the day the spreadsheet was created. Without the full history there is no way to check whether last year's spend changed anything.
Why it works this way
What survives of the spreadsheet once analysis is automatic
What survives is what a spreadsheet is genuinely good at: your own notes, deadlines, assigning tasks to people. The CSV export from every analysis gives you reviews already classified, with date, platform and sentiment — ready to paste into a spreadsheet that remains yours. The copying disappears; the work stays.
Common questions
Can I import my existing spreadsheet?
There is no such feature and honestly you do not need one: the analysis pulls the full review history from the platforms anyway, including everything in your spreadsheet. Your own notes stay where they are.
Will I get data in a format I can paste into a spreadsheet?
Yes. The CSV export contains every review in the corpus with date, platform, rating, categories, sentiment and the full source text.
How is the model's classification better than mine?
It is not smarter — it is consistent. The same review gets the same category regardless of the day and the person, and that is the only condition under which percentage shares mean anything.
See also
Be first in when we launch
Join the waitlist. We'll write once, on launch day, with a link to create your account.