Method
How a tool gets its verdict, and how we count the saving the guarantee rests on.
Three questions per tool
Use
What share of the paid seats was really used last month.
Impact
What stops the day the tool goes.
Overlap
Whether something else in the stack already does the job.
Two things fail a tool at once: a zombie subscription under 30% use, and a thin layer around a public AI model that you can also use directly. In doubt it is a Kill: a tool you have to defend does not earn its place.
How we count savings
| Kill | The full cost per year. |
|---|---|
| Switch | The cost per year minus the cost of the alternative. If the alternative is already in your plan, that is zero. |
| Keep | The yearly cost of the unused seats. |
The total is at the bottom of the report. That amount is compared with your fee, and that is the guarantee. It counts what the report finds, not what you cancel afterwards. We calculate costs from your invoices or contract prices, and only if you have none from public list prices.
Try it with one tool
Run one tool