Nobody decides to spend more on fixed costs. It happens by accumulation: a tool added for a project that ended, a plan that upgraded itself on usage, a retainer that renewed because cancelling it was a conversation nobody wanted, an insurance premium that rose quietly at renewal. None of it is a decision and all of it is permanent until somebody looks. This is a ninety-minute routine, once a quarter, that prices every fixed line in the only unit that makes the question concrete, which is orders, and then cancels whatever nobody is willing to defend.
- Price every line in orders: at €44,68 of contribution, €100 a month is 27 orders somebody has to sell.
- Sort by size, not by irritation. The three biggest lines matter more than the six annoying ones.
- Ask each line for its defender. A cost nobody will name themselves against is a cost nobody is using.
- Diarise renewals so the next increase is a decision rather than a discovery.
Convert everything into orders
A monthly figure invites a shrug and an order count does not. A €100 subscription is 27 orders at the example store's contribution per order, and 27 orders is a day of trading for many stores, which is exactly the framing that makes a decision happen. Do the conversion for every line before discussing any of them.
It also puts the lines in the right order of importance. Software bills are the ones that annoy people, and rent, salaries and agency retainers are the ones that decide the month; a review that spends an hour on subscriptions and five minutes on the retainer has optimised the wrong end.
Find the two failure patterns
The first is the orphan: a tool bought for a project, a person or a season that has ended, still billing because nobody owns cancelling it. The test is naming a defender. If no person in the business will say the words that they use it and would notice its absence, cancel it and see who complains, which is usually nobody.
The second is the silent upgrade: plans that scale with seats, usage or revenue and step up without anybody approving the step. Those need checking against what they were last quarter rather than against what they should cost, because the increase is invisible in a monthly total that was already normal.
| Question | A bad answer looks like | What to do |
|---|---|---|
| Who uses it? | nobody will name themselves | cancel it |
| What did it cost last quarter? | less, and nobody noticed | check the plan tier |
| When does it renew? | unknown | diarise it now |
| What breaks without it? | a vague answer about the future | downgrade and see |
Spread it, then judge it
Fixed costs are consumed daily even when they are billed annually, so a review that looks at invoice dates rather than daily burden will mis-rank everything: a €1.200 annual tool looks smaller than a €150 monthly one until both are spread across the days they cover. Proration is what makes the comparison fair, and it is also what stops one bad Tuesday every year when the insurance bills.
What to do with what you find
Cancelling is the obvious move and it is not always the right one. Three softer options usually recover most of the money with none of the disruption: downgrade a tier that was upgraded for a peak that has passed, move an annual renewal to the month that suits your cash rather than the month it happened to start in, and consolidate two tools that overlap into whichever one somebody actually opens.
Whatever you change, put the new total straight back into the two numbers it moves: cutting the block lowers the break-even order count and widens the cushion, which a margin of safety calculator will show in days of trading. Seeing that effect immediately is what makes the routine survive to the next quarter.
The line worth protecting
One warning about doing this too well. A quarterly cost review is easy to turn into a ritual of cutting, and some fixed costs are the reason the business works: the warehouse that ships on time, the accountant who prevents expensive mistakes, the tool that saves a day a week. The question is never whether a line is large, it is whether it earns its order count, and a line that clearly does should be left alone quickly so the meeting can spend its time on the ones that do not.
The ninety minutes
- Export every recurring payment from the bank and the card, not from memory. The forgotten ones are the point.
- Convert each into a monthly figure and then into orders at your own contribution per order.
- Sort by size and work down until the numbers stop mattering, rather than working through the list alphabetically.
- For each of the top ten, name a defender, a renewal date and what breaks without it.
- Cancel what fails, diarise the rest, and put the total back into your break-even and your cushion so the effect is visible next month.
Why quarterly and not annually
Annual reviews find annual problems: by the time twelve months have passed, an orphaned tool has cost a year and a silent upgrade has become the new normal. Pairing the review with the month end close checklist keeps it on a calendar rather than on somebody's conscience. A quarter is short enough that the creep is still recognisable as creep, and long enough that the routine does not become an irritation. Ninety minutes, four times a year, against the block that decides how many orders your month needs before it earns anything.