Cost control & analyticsIn production

Which people have not used a paid subscription in 30 days? One page answers it.

A report of paid seats with no recorded usage, priced per seat. Tools nobody can observe are listed separately, so "unwatched" is never mistaken for "unused".

30daysdefault idle window, with 60 and 90 one click away
6usage signal tiers, from per-login data to nothing
Idle seats report listing people who have not used a paid tool in 30 days, grouped by tool
Rebuilt interface · illustrative data

The business problem

The agency pays for 30 to 40 software tools. Nobody could say which paid seats were unused. Vendor reports cover only part of the portfolio, and a seat cannot be cut safely without evidence that its holder has stopped using it.

What changed

Yash built an internal app with one job. It reads sign-in records from the company's Google Workspace, last-active data from vendor connectors and, where neither exists, a managed browser extension. The home page lists every person and paid tool pair with no usage inside the window. It can be grouped by tool or by person, and exported as a CSV.

Before & after

AreaBeforeAfter
Who is idleUnknownA dated list of seats with no usage in 30, 60 or 90 days
LeaversSeats can stay assignedPeople who left the company are flagged on the seat they still hold
Unwatched toolsSilence read as non-useListed apart, with the next step to start seeing usage
Flat-priced plansA per-seat saving impliedNo per-seat figure shown, because removing one person saves nothing

The business value

The value is evidence for renewal and seat decisions. Each idle seat carries its monthly cost, and the page totals what could be reclaimed. The report only counts tools whose usage can truly be observed.

How to read the return

The app is live, but no savings figure has been measured yet, so none is claimed. The figures on the screens are made up.

Why the system is dependable

The split between observed and unobserved tools is enforced in the database, not just on screen, and a test pins it. The browser extension counts one number per approved work tool per day. It never records page addresses, titles, time spent or any other site, and sites outside the list are discarded inside the browser. The project also includes an employee privacy notice. The repository has 28 test files, and a daily scheduled sync keeps usage current.

A closer look at the build

Paid tools with no usage source, each with the next step to start seeing usagePaid tools we cannot see (full size, opens in a new tab)

Rebuilt screens in the style of the real product. All tools, people and amounts shown are fictional.

Could this work for your team?

If you pay for many seats and cannot say who uses them, this approach gives you a defensible list for renewals. The honest part is knowing which tools you cannot see yet.

Talk through a similar project
Keep exploring

More case studies

All case studies