Software spend grows quietly. Nobody decides to spend six hundred a month on subscriptions. It accumulates, one twenty-nine at a time.
TL;DR List every subscription and what it costs annually. Value time saved at your loaded hourly cost, not the sticker rate. Kill anything with no owner. Consolidation usually beats optimisation.
The audit
Start with the bank statement, not with memory.
Twelve months of card and bank transactions, every recurring charge.
For each, record
- Name and what it does.
- Monthly and annual cost.
- Who owns it.
- Who actually uses it, and how often.
- What breaks if it is cancelled.
- Renewal date and notice period.
Findings that come up nearly every time
- Subscriptions for tools nobody has opened in months.
- Two tools doing the same job.
- Per-user plans with seats for people who left.
- Annual plans auto-renewed nobody reviewed.
- A trial that converted and was never noticed.
- A tool bought for one project, still running.
Cancel the obvious dead weight immediately. That usually pays for the time spent on the audit several times over.
Valuing time saved
The most common mistake is valuing time at the hourly rate you charge.
Use loaded cost, not billing rate.
Salary plus employer taxes plus benefits plus overhead allocation, divided by actual working hours.
And be honest about whether the saved time is recovered. Saving a technician twenty minutes a day only creates value if that twenty minutes becomes billable work or reduces overtime. If it becomes twenty more minutes of standing around, the value is zero.
Two categories, valued differently
| Type | How to value |
|---|---|
| Time that becomes billable work | Full contribution margin |
| Time that reduces overtime | Overtime rate |
| Time that reduces headcount need | Loaded cost |
| Time that just makes the day nicer | Real, but not financial. Do not put a number on it |
The last row matters. Some tools are worth keeping for morale and retention reasons and should be justified that way, honestly, rather than with a fabricated productivity figure.
The ROI calculation
Keep it simple and conservative.
Annual benefit minus annual cost, divided by annual cost.
Annual cost includes
- Subscription.
- Implementation time, amortised over a realistic life.
- Training time.
- Ongoing administration.
- Integration maintenance.
Annual benefit includes
- Time saved, valued as above.
- Revenue enabled. Jobs won that would have been lost.
- Errors avoided, valued at the cost of the error.
- Cost replaced. What it stopped you paying for.
Round everything down on benefits and up on costs. A tool that survives a pessimistic calculation is genuinely worth keeping.
Revenue-enabling tools are different
Some tools do not save time. They win work.
Value these on outcome, not efficiency
- Missed-call text back: leads recovered times conversion rate times average job value.
- Quote follow-up automation: incremental acceptances.
- Review generation: harder to attribute, but visible in enquiry volume over time.
These frequently return far more than the efficiency tools and get scrutinised less because their cost is smaller.
Measure them properly. Turn one off for a fortnight if you can do so without customer harm, and compare. It is the only clean test available at small scale.
Cost per processed lead
A useful comparative metric across your whole stack.
Total monthly software cost attributable to lead handling, divided by leads processed.
Then compare against
- Cost per lead acquired, from marketing.
- Value per lead converted.
If your software cost per lead approaches your acquisition cost per lead, the stack is oversized for the volume it handles. That is a common finding in businesses that bought for the scale they planned rather than the scale they have.
Redundancy audit
Map tools against functions.
| Function | Tools currently doing it |
|---|---|
| Scheduling | Field service app, calendar, booking tool |
| Customer messaging | Email, SMS tool, social inboxes, chat widget |
| Payments | Processor, accounting, field app |
| Storage | Two cloud drives, plus the field app |
Wherever a row has more than one entry, ask why. Sometimes there is a good reason. Often it is history.
Consolidation benefits beyond cost
- One source of truth for data.
- Fewer logins and less training.
- Fewer integrations to maintain.
- Reporting that works without stitching.
Consolidation usually beats optimisation. One adequate platform doing six jobs beats six excellent tools with five bridges between them, at this scale.
The kill list
Cancel anything that meets any of these
- No named owner.
- Not opened in 90 days.
- Duplicates a function you already pay for.
- Was bought for a project that ended.
- Nobody can explain what it does.
Before cancelling
- Export your data. Always, first.
- Check what depends on it.
- Check the notice period.
- Tell whoever might be affected.
- Diarise the actual cancellation date.
Set a calendar reminder two weeks before every renewal, permanently. Auto-renewal is how most of this accumulates.
Buying discipline going forward
Before adding anything
- What specific problem, and how often does it occur?
- Can an existing tool do it, adequately?
- Who owns it?
- What is the annual cost at our size in two years?
- How do we get our data out?
- What is the review date?
Set a review date at purchase, six months out, in the calendar. A tool that has not proved itself in six months rarely does in twelve.
Pull twelve months of recurring charges this week and list them against what they do. Nearly every business finds something to cancel in the first pass, and that finding funds the rest of the exercise.
Need a pro to audit it? [BOOK A CALL]