The Subscription Audit

No small company decides to spend a meaningful monthly sum on software. It decides fourteen times to spend a small one, across three years, each decision perfectly reasonable at the time, and it never decides to look at the total.

The pattern is always the same. Somebody needed a thing. They found a tool. The entry plan was cheap. The team grew and per-seat pricing multiplied. The tool was partly replaced by another one, and nobody cancelled the first because an old report still lives there.

The audit, in four steps

1. Extract from the statement, not from memory. Take twelve months of corporate card and bank statements and filter everything recurring. The real list is always longer than the remembered one — annual charges disappear from memory, and they are the large ones.

2. Mark who uses it, not who has access. For each tool, how many people logged in over the last thirty days. Most services show this in the admin panel. Paid seats for people who have left the company are the most common finding and the easiest to fix.

3. Sort into four categories. This is the step that produces the decision:

• Infrastructure — if it stops, the operation stops. Hosting, database, email, telephony. You do not cut these; you negotiate them.
• Workflow — the team uses it daily. Replacing means retraining. Cut only with a migration plan.
• Convenience — solves something a cheaper tool or a manual step would also solve. Nearly all the savings live here.
• Zombie — nobody has logged in for sixty days. Cancel immediately.

4. Cancel the zombies the same day. Do not hold a meeting about it. If nobody has opened it in two months, nobody will miss it in two more.

Where the money hides

• Seats for people who left. Almost every company has them. First place to look, fastest return.
• A plan above actual usage. Tools push the upper tier for one feature you used once. Check real consumption before renewing.
• Two tools doing the same job. This happens when two people solve the same problem in different months. Chat, storage and task management are the usual offenders.
• Annual value paid monthly. Anything you are certain to keep is usually materially cheaper on an annual plan. The reverse also holds: never go annual on something still being trialled.
• Forgotten add-ons. Extra storage, an additional user, a premium integration. They accumulate quietly and never come up in conversation.

What we recommend — including the contrarian part

The market reflex is to sell you a subscription-management tool to manage subscriptions. For a small company that is almost always wrong: you would be adding a line to the problem you are trying to solve.

1. A spreadsheet, reviewed quarterly. Columns: tool, category, monthly cost, active users, renewal date, owner. Fifteen minutes a quarter beats any software. It is the argument from Your Competitor Is a Spreadsheet, turned on yourself.
2. One card dedicated to subscriptions. Separating it turns the audit into reading a statement instead of doing accounting archaeology.
3. One owner per tool. A person's name, not a department. A tool with no owner is a tool nobody cancels.
4. A reminder thirty days before each annual renewal. That is the only window in which you have negotiating power — and a renewal discount usually exists if you ask for it.

Where a tool does earn its place: above a few dozen subscriptions, or when several people can buy without passing through a single point. Below that, the spreadsheet wins.

The cuts not worth making

Being specific about this matters as much as the list of what to cut.

Do not cut backup. Do not drop professional email for a free account. Do not move hosting somewhere cheaper without understanding the migration cost. Do not cancel a workflow tool without a plan — the productivity lost during adjustment frequently exceeds a year of the saving.

A subscription audit is about removing what is not used, not about paying the minimum possible for everything.

A company that optimises every line for lowest price ends up paying somewhere else, and that somewhere is usually time — which is the most expensive resource a small operation has.