← Blog
  • productivity
  • tools
  • operations

How many tools do you actually need? A 90-minute audit

7 min read

Most small businesses pay for three tools doing the same job badly. A timed audit to see what you use, what overlaps and what you can cancel today.


Open your bank statement and look for the small charges: €12, €29, €9.99, €45. None of them hurts on its own, which is exactly why they have been sitting there for eighteen months. But the expensive part is rarely what you pay. It is that three of those tools do the same job badly and nobody on your team knows which one to trust. What follows is a ninety-minute audit, on the clock, that ends with a short list of tools you actually use and a single page saying where everything lives.

The €29 a month is not the problem

Add up your subscriptions, get an uncomfortable number, and the instinct is to cancel the most expensive one. That is the wrong cut. Having too many tools costs you in three other currencies.

The first is the "where was that again?" tax. Every time someone hesitates between the inbox, the spreadsheet and the app, you lose thirty seconds and a little faith in the system. Thirty seconds, forty times a day, across three people, is a morning a week that never shows up on an invoice.

The second is split data. If the same customer lives in WhatsApp, in a spreadsheet and in your invoicing software, you do not have three sources — you have none. No copy is complete, none is authoritative, so the real version stays in your head.

The third is silent maintenance. Every tool comes with its own user admin, its shared password that should not be shared any more, its invoice to reconcile and its quiet price rise. Nobody owns that work, which means you do it.

There is one signal that never lies: if a new hire needs more than one page to learn where things go, you have too many tools.

Ninety minutes, four blocks

Block out a Tuesday morning, not late Friday. You need your bank statement, your inbox and an empty spreadsheet.

Minutes 0–20: the real list, not the one in your head

Do not do this from memory. Pull twelve months of transactions from the account and the cards, then add PayPal and the app store on your phone. Twelve months, not three — annual plans hide there, and those are precisely the ones that renew themselves without anyone noticing.

One row per tool, four columns: what it is, annual cost, who pays for it, who actually uses it. That last column holds a person's name or stays empty. "The team" is not an answer.

Then add the free tools that hold data. They are the expensive trap on this list: they never appear on a statement, but that shared folder or that free board weighs just as much as a paid seat the day you try to tidy the business up.

Minutes 20–50: one row per job, not per tool

This is the whole trick. Flip the sheet around: instead of listing tools, list the jobs your business does. There are usually ten or twelve, and they are almost always the same ones — get leads, reply, schedule, quote, deliver, chase payment, invoice, store files, coordinate the team, do marketing.

Next to each job, write the tool that does it today. The overlaps surface on their own, and there is nothing to argue about:

  • A job with two or three tools next to it is an overlap.
  • A tool that appears against no job at all has been a direct debit for months.
  • A job with no tool is the one you have been doing by hand without noticing.

That third case is usually the most valuable find of the morning, and it saves no money at all. It saves hours.

Minutes 50–75: four decisions, no nuance

Every tool gets one letter. No long debates, no "it depends".

  1. K — Keep. It is the official tool for its job. Done.
  2. M — Merge. It duplicates something else; migrate to the official one and set a date.
  3. C — Cut. Nobody uses it, or its job no longer exists.
  4. P — Park. Genuine doubt: thirty days watching whether anyone opens it, then decide again.

The rule that makes this work is deliberately uncomfortable: each job gets exactly one official tool. Not two "depending on the case". If you truly need two, write the boundary between them in one sentence, and make it a sentence someone who started last week could follow.

A tool can survive without being the best one. It cannot survive without being the only one for its job.

Minutes 75–90: act before you stand up

Whatever you do not cancel today does not get cancelled. But do it in this order: export first (nearly every tool lets you pull a CSV or a zip, and the moment you cancel that data stops being one click away), then cancel, then put the renewal dates of the survivors in the calendar with a reminder a week ahead.

Finish with the only artefact your team will actually read: a ten-line list, "job → tool", sent on whatever channel you use. Without it, someone reopens the tool you just killed within three weeks.

The five overlaps you will almost certainly find

Do this exercise across a handful of small businesses and the same five keep showing up:

  • The customer in three places: the conversation in WhatsApp, the details in a spreadsheet, the history in the invoicing software.
  • Two calendars that do not talk, usually because someone started their own when the shared one got crowded.
  • Tasks in four: phone notes, a board that lasted two weeks, starred emails, and your memory.
  • Duplicate storage: Drive and Dropbox side by side, plus the email attachments that never made it to either.
  • The tool from a project that ended eight months ago and still bills.

None of these is fixed by buying something new. They are fixed by deciding which of the things you already pay for is the official one.

When not to cut

You can do this audit badly, and it is almost always by cutting too much.

Leave anything with retention obligations alone. Invoicing, accounting and anything holding supporting documents does not get cancelled to save fifteen euros — check with your accountant first, and if you do migrate, migrate with a verified export in hand.

Respect the odd tool that only one person uses, if that person is fast in it and the work comes out right. Standardising is worth something, just not enough to break someone who already works well. Where two people do the same job in different tools, though, pick one.

The last one is obvious and still gets ignored: do not migrate during your busy season. Moving data mid-campaign is how orders go missing. Put the date in the quiet month and write it down.

Do this before you automate anything

There is a practical reason to run the audit now rather than "when there's a gap". Automation multiplies whatever it sits on. An agent that has to check three places to know whether a customer has paid will give you three different answers, and that is not the agent's fault — it is the map you handed it.

Once every job has one official tool, delegating stops being a project. Whether it goes to the person starting in September or to an agent handling follow-ups, reminders and quotes, the instruction fits in a sentence, because there is one place to read from and one place to write to.

Run it again in six months. The second pass takes forty minutes and turns up less, which is exactly the sign that it is working.

If you finish with fewer tools but still nobody to do the dull work behind them, that is where an operations agent fits: it works inside the tools you decided to keep, and nothing goes out until you approve it.