Ask the owner of almost any small business whether a piece of software has been worth it, and the answer is usually a confident "yes, I think so". Ask for the number behind that answer and the conversation changes. Nobody wrote down how long the job took before the tool arrived, how many enquiries were converted, or how many hours of capture work the new system was meant to remove. The software was bought, switched on and absorbed into the monthly debit orders, and the question of whether it earned its place was never asked in a form that could be answered.
A tolerated habit: judging software by feel
This is not carelessness. It is how most purchasing decisions are made. A tool is chosen because a demonstration looked impressive, a competitor uses it, or a frustrated manager finally got approval. Success is then assumed. Because nobody recorded the starting position, nobody can later prove improvement, and nobody can prove failure either. The tool simply stays.
The habit is tolerated for three reasons:
- Measurement feels like extra work. The team is busy and the implementation is already late. Taking a baseline looks like a luxury.
- Benefits are spread thinly. Ten minutes saved per person per day does not show up as a line in the accounts. It disappears into a longer lunch, a shorter queue or a quieter inbox.
- Nobody is accountable for the outcome. The person who bought the tool has moved on to the next problem, and finance only sees the invoice.
Mellorca has written about subscriptions nobody can see and about the practical method in measuring whether software pays for itself. This article is about the consequence of skipping that step: spending that keeps growing with no evidence either way.
How it shows up day to day
You can recognise the pattern without any special tooling. Typical symptoms include:
- Managers describe a system as "essential" but cannot name the task it made faster.
- The same job is still done in a spreadsheet alongside the new system, "just to be safe".
- Licences are bought for everyone, yet only a handful of people open the tool in a normal week.
- Renewal arrives and the only question asked is whether the price has changed, not whether the value has.
- Complaints about the tool are common, but compliments are vague.
Each of these points to the same gap. There is no agreed definition of what "worth it" means, so every opinion is equally valid and none can be tested.
Where the money actually goes
Unmeasured software costs the business in several separate ways, and the subscription fee is often the smallest.
- Technology waste. Licences and plans that nobody uses properly keep being paid for. Our article on unused software licences covers the direct fee; the point here is that without a baseline you cannot even tell which spend is waste.
- Labour cost. If a tool was meant to remove manual work and did not, the staff time is still being paid for. If it did remove work, but the freed time was never redirected, the saving is invisible and quickly reabsorbed.
- Opportunity cost. Money and attention committed to a tool that is not working cannot go to one that would. Poorly measured projects also crowd out better ones.
- Revenue left unclaimed. A tool bought to speed up quoting or follow-up may be delivering real gains, but if nobody tracks response times and win rates, the business cannot see where more investment would pay off.
A transparent way to estimate the exposure
You do not need precision to see whether the problem is worth your attention. Here is an illustrative example, using made-up round numbers for a South African services firm of about 25 people. These are assumptions for the calculation, not benchmarks.
| Item | Assumption |
|---|---|
| Business applications paid for each month | 14 |
| Average monthly cost per application | R3 000 |
| Annual software spend | R504 000 |
| Applications with a recorded before-and-after measure | 2 |
| Applications where nobody can say what changed | 12 |
On those numbers, roughly R432 000 a year is being spent on tools whose contribution is unproven. That does not mean the money is wasted. It means the business is trusting twelve purchases it has never checked. If a quarter of them turn out to be marginal, that is around R108 000 a year of recoverable spend, before counting any labour. Replace the assumptions with your own figures, and the exercise takes an hour.
Why it is hard to fix after the fact
The difficulty is not analysis; it is the missing starting point. Once a tool has been live for a year, nobody remembers how long the old process took. Memory favours the new system because it is familiar and the old pain has faded. Cost-benefit arguments made later rely on recollection and tend to favour whoever is most confident.
This is why guidance on adopting technology puts so much weight on setting up measures before the change. Microsoft's Cloud Adoption Framework planning guidance, for instance, tells organisations to define success metrics and establish a cost baseline, and to re-evaluate estimates when actual figures drift from the plan. That advice is written for cloud migrations, but the principle applies to any business software: a baseline is the thing that turns an opinion into a comparison.
The commercial trigger point
A few events should make the question urgent:
- A renewal is coming and the vendor has announced a price rise. Our article on automatic renewals without a value review explains why this is the cheapest moment to ask hard questions.
- The software budget has grown faster than headcount or turnover for two years running.
- A manager asks for yet another tool to fix a problem that the last tool was supposed to fix.
- The business is planning a major investment, such as automation or a new platform, and needs to know what the current systems really deliver.
What better looks like
Better does not mean a heavy measurement programme. It means a few habits applied consistently.
- Write down the job before buying. One sentence: what task, done by whom, how often, taking how long, producing what error rate or delay.
- Pick one or two measures. Hours per week, days to invoice, enquiries answered within a day, rework incidents. Choose measures staff already touch.
- Record the baseline for two to four weeks. Rough tallies are enough. A tick sheet beats nothing.
- Name an owner. A person, not a department, who answers for whether the tool delivered. Our article on unclear ownership of technology spending shows what happens without one.
- Set a review date. Ninety days after go-live, and again before each renewal.
Using what the platforms already tell you
Usage evidence is often already available. Microsoft documents usage reports in the Microsoft 365 admin centre that show who uses a service heavily and who barely uses it and might not need a licence. The reports cover the last 7, 30, 90 and 180 days, and user names are hidden by default to support privacy requirements. Google Workspace has comparable reporting in its admin console, with the available reports depending on your edition, and most other business platforms offer something similar. Usage is not the same as value, since someone can open a tool every day and still be fighting it, but low or falling usage is a cheap early warning that a purchase is not landing.
Pair usage data with a short conversation. Ask three people who use the tool: what does it let you do that you could not do before, what still takes too long, and what would you miss if it went away tomorrow? Honest answers to the third question are very revealing.
Practical diagnostic steps
- List every paid application with its monthly cost, its owner and the job it was bought for.
- Mark each as "measured", "plausible" or "unknown". Be honest. Most will land in the last two.
- For the three most expensive "unknown" items, reconstruct a rough before-and-after from invoices, ticket counts, turnaround times or staff estimates.
- Check usage data against licences held.
- Decide for each: keep and measure, change plan, or retire at the next renewal. Our guide to rationalising software without breaking the business covers the safe order of doing this.
Secondary consequences
Unmeasured technology spending also distorts decisions elsewhere. A business that cannot show which tools work will struggle to defend its technology budget when money is tight, and it may cut the wrong thing: the cheap tool that quietly saves hours, rather than the expensive one nobody uses. It also makes staff cynical. When every new system arrives with promises and none is ever reviewed, people stop investing effort in learning the next one.
The bottom line
Software is a cost that is supposed to buy time, accuracy or revenue. When no one measures whether it does, the business pays the fee, pays for the manual work that continues around it, and pays again in lost opportunity when better ideas cannot be funded. The remedy is modest: a baseline, a small number of measures and a named owner, applied before the next purchase and before the next renewal. A business that can say what each tool is worth is in a far stronger position to decide what to keep, what to change and where to invest next.
Sources and further reading
- Microsoft 365 Admin Center Usage Reports Overview - Microsoft 365 admin (last verified 10 October 2026)
- Plan - Training
- Monitor usage & security with reports - Reports & monitoring - Google Workspace Help (last verified 10 October 2026)
Related Mellorca reading
- Accepting Every Price Rise: How Unreviewed Renewals Compound Year After Year
- Paying for Tools You Already Own: The Cost of Underused Software Platforms
- Software Bought Without Approval: The Financial and Security Cost of Shadow IT