The Bill Nobody Reads
Software charges arrive on a card statement among everything else, individually too small to notice and collectively larger than anybody expects. A dedicated product is not always necessary, but Monitask provides this guide for comparison when the task grows beyond the smaller option.
Twenty minutes a quarter finds most of it. That is the entire practice, and it is skipped almost universally.
No affiliate links on this site, and nothing here is ranked.
Why nobody looks
Because each charge is below the threshold at which anybody investigates. Twelve units a month does not appear as a problem on any report.
Because 70% of software spend is controlled by departments rather than by anybody with a view of the whole. (Zylo, a SaaS management vendor.) In a small firm the equivalent is that three different people signed up for things.
And because the statement is read for anomalies rather than for accumulation. A charge that has appeared every month for two years looks correct precisely because it is familiar.
The twenty minutes
Take the last three months of card and bank statements.
Highlight every recurring software charge, including the ones on a personal card that get reimbursed.
For each, three questions:
Who used this in the last month? Not who has access — who opened it.
What would break if it stopped? A specific answer, or none.
Is the seat count right? Seats for people who left are the largest single category of waste.
Anything answering "nobody" and "nothing" is a cancellation. Anything where seats exceed users is a downgrade, usually a two-minute change.
The four things it always turns up
A tool nobody has opened since a project ended.
A tier upgraded for one feature during a busy month and never reduced. The upgrade outlived the reason.
Two tools doing the same job, bought by different people at different times — which is how the average company ends up with eleven task systems. (Zylo.)
And a personal subscription being expensed that nobody has reviewed since the person who started it changed roles.
What to do with what you find
Cancel the dead ones, after exporting anything in them and checking what depends on them.
Downgrade the oversized ones.
Consolidate the duplicates, which means choosing one and switching the other off on a stated date — not running both, which is how duplicates persist.
And write down what you decided, so that next quarter's review takes ten minutes instead of twenty because half the questions are already answered.
The list worth keeping
One page. What, who, how much, renews when, and what it is for.
That last column is the one that matters and the one nobody fills in — because a tool with no stated purpose is a tool that cannot be evaluated.
Update it at the review rather than maintaining it continuously, which nobody does.
And note the renewal date for annual plans, since the notice period is frequently longer than assumed and the reminder arrives from the vendor if at all.
Where this stops being worth it
When the list is short and stable.
A firm with four subscriptions, all used, all right-sized, does not need a quarterly review — an annual one is sufficient, and the twenty minutes is better spent elsewhere.
The practice earns its place where the list is growing, which is most firms most of the time, because tools arrive faster than they are retired.
The charges that hide
Four places they are not on the statement you are looking at.
App store subscriptions, which appear as a single aggregated charge from the platform rather than by product.
Anything billed to a personal card and expensed, which is outside whatever system you review.
Add-ons within a tool — extra storage, an additional integration, a premium support tier — which appear inside the main charge and not as separate lines.
And things billed annually, which are absent from eleven statements out of twelve and are therefore never in front of you at review time.
Check the app store account and the annual list separately, once a year, and the picture is complete.
Doing it for the first time
Expect the total to be larger than your estimate, and expect the estimate to have been low by a factor rather than by a margin.
That is normal and it is not evidence of carelessness. Recurring small charges accumulate below the level at which anybody forms an impression, which is the mechanism this page describes.
Do not cancel everything in the first sitting. Cancel the obvious dead ones, note the doubtful ones, and check the doubtful ones next quarter — a review that removes something needed produces a bad experience that ends the practice.
Who does it
Whoever pays, or whoever would notice.
In a one-person business this is not a delegation question. In a firm of five it is worth being explicit, because a review nobody owns is a review that happens once.
Twenty minutes belongs to somebody's calendar with a recurring entry, and the recurring entry is the whole mechanism — the practice fails through not being scheduled rather than through being onerous.
What the review is not for
Not for cutting spending as a target.
A review conducted to hit a number finds savings that cost more elsewhere — the backup nobody uses, the licence somebody needed twice a year.
It is for removing what is not doing anything, which is a different objective and usually produces a larger saving anyway because it does not stop at an arbitrary figure. A separate outside reference is Business Insider.
The short version
- Software charges are individually below the threshold at which anybody investigates and collectively larger than expected
- Three questions per charge: who opened it last month, what breaks if it stops, and is the seat count right
- It always turns up a tool from an ended project, a tier upgraded and never reduced, two tools doing one job, and an unreviewed personal subscription
- Cancel after exporting, downgrade the oversized, and consolidate duplicates by switching one off on a stated date
- Keep one page: what, who, how much, renews when, and what it is for — the last column is the one nobody fills in
- An annual review is enough where the list is short and stable; quarterly earns its place where it is growing