There is a point in a growing firm where the answer to every constraint becomes a purchase. Capacity is tight, so the plan is another tool. A report takes three days to produce, so the plan is a platform that produces it. The reasoning is sound and the spend is affordable, and it skips the step that pays better than the purchase does.
That step is counting what the firm already pays for.
Firms carry capacity they have already bought and are not using. It is not waste in the sense of money thrown away, because the money bought something real. It is capacity switched on halfway, configured for a firm two sizes smaller, or handed to somebody who has since left. Switching it on is faster and cheaper than buying the next thing, and it arrives without a migration, without a new login for the team, and without the three months where nothing works properly.
What counting it involves
A paid-for capacity audit is a short piece of work with a long payback. It asks one question of every recurring cost and every system in the firm: what did we buy this to do, and is it doing it? The answers sort into four piles. Using it. Using part of it. Not using it. Did not know we had it.
The last two piles are where the money is, and they are also where the fastest operational gains sit, because the thing has already been paid for, chosen and installed. What remains is a configuration job and a decision.
Five places the capacity sits
Across engagements the same five keep appearing.
A seat nobody sits in. Per-user licences are the easiest of the five to find and the most commonly missed, because the invoice arrives as one figure rather than as a list of names. Seats stay live for people who left, for contractors who finished, and for a trial that was set up with five and needed two. Five seats at $40 a month is $2,400 a year against a firm that never onboarded three of them.
A feature that came with something already bought. This is the expensive one. A firm buys a specialist tool for document automation, or time capture, or client onboarding, and the practice management system it has paid for since 2022 has a module that does the same job. The module was never configured, because configuring it was a project and buying was a decision. The firm now pays twice, and the two systems hold the same client list with slightly different spellings.
Work the firm has already paid to have done once. A scope document, a client onboarding pack, a fee schedule, a report format. Somebody built it properly, billed the hours to build it, and it now sits in one matter folder. The next person who needs it builds it again from nothing, because no step turned a one-off into an asset. The firm paid twice for the same artefact and owns neither version as a template.
A skill the firm trained and never used. Somebody was sent on the course, passed the certification, and came back to the same work they were doing before. The training was real and the capability is real, and there is no work pointed at it. This one costs a person as well as money, because being trained for work you are never given is the shape of a resignation six months out.
A renewal nobody has made a decision about. Annual subscriptions renew silently, and a renewal is a purchase that nobody signs. Plenty of firms are carrying tools that made sense for a problem that no longer exists, renewed every October since, with no moment in the year at which a person is asked whether to keep it.
Repair before replace
When something in the firm is not working, the choice sits between fixing what is there and buying something new. Three questions decide it, and they take an hour.
Is it broken, or is it unconfigured? A tool nobody set up properly behaves the same as a tool that does not work. The difference matters, because one of them is a replacement project and the other is an afternoon with the right person and the settings page.
What does the switch cost in time, not in licence fees? The replacement's price is the visible number. The real cost is data migration, re-training, the months of two systems running side by side, and the jobs that fall between them. We put a figure on those hours before comparing the two subscriptions.
Who owned the old one, and will they own the new one? A system with no named owner drifts out of true whatever it cost. If the answer to who owns it reads the same after the purchase as it did before, the new tool inherits the problem that finished the old one.
Replacing is sometimes the correct call. A tool built for a firm a third of this size, or one whose vendor has stopped developing it, is a replacement. The three questions exist so the decision has evidence under it before anyone signs.
The exercise: a renewal list
This takes an afternoon and it finds money in most firms that run it.
It starts with twelve months of card and bank statements. Every recurring charge goes on the list: the name, the monthly figure, the annual figure and the renewal date. The small ones go on too, because the small ones are where the forgotten items live.
Then each line gets one of four marks. Using it. Using part of it. Not using it. Did not know about it. The team answers for their own tools, because the person paying the invoices is rarely the person using the tool.
Finally, each renewal date gets a diary entry two weeks ahead, addressed to a named person. That entry is the decision the firm has not been making.
What the good version looks like
A firm doing this well knows what it spends on systems as one figure, and can say what each line buys. Renewals arrive as a decision on a calendar rather than as a charge on a statement. New capability gets checked against what is already installed before a demo is booked, which shortens the list of tools rather than lengthening it.
The second-order effect is the one firm owners notice. Fewer systems means fewer places a fact can live, fewer logins for a new starter, and fewer of the small reconciliations that eat a Thursday afternoon. Reusing and repairing what is already owned tends to leave a firm simpler than it found it.
Where we start
We build the renewal list before the next purchase decision arrives. The lines marked not using it and did not know about it get cancelled or configured within a fortnight. Then the single most expensive line marked using part of it gets a proper look at what the rest of it does.
A firm that knows what it already pays for buys less, and the things it does buy arrive into a business that has room for them.