A cost-benefit analysis sets out what a decision will cost and what it will return, so the comparison is explicit rather than a feeling.
Costs that get left out
Ongoing costs. The purchase price is rarely the cost. Insurance, maintenance, training, licences and replacement all follow.
Time. Somebody has to implement it, learn it, and run it. In a small business that somebody is frequently the owner, and their time is the scarcest input — see opportunity cost.
Disruption while changing over, which is real and temporary and almost never budgeted.
Benefits that get overstated
Benefits assumed at full effectiveness from day one. Most arrive gradually and some never fully arrive.
Benefits counted twice — a tool that "saves ten hours a week" only saves money if those hours are then billed or removed from the payroll. Otherwise the hours are simply spent differently.
Putting numbers on the awkward parts
Some benefits resist measurement: less risk, better information, a calmer week.
They are still real. The honest approach is stating them plainly alongside the numbers rather than inventing a figure, and being clear which part of the case rests on them.
The comparison that matters
Against the next best use of the same money, not against doing nothing — see return on investment.
Revisiting it
A decision made on a case is worth checking against what actually happened, once. It is the only way the next estimate improves, and it is the step almost always skipped.