Count the cost after launch

The cheapest invoice can belong to the most expensive arrangement. A product may cost little while requiring an operations lead to reconcile its records every week. A custom build may avoid seat charges while creating more specialist maintenance than the business expected.

Cost follows the work, including work that never receives an invoice. A useful comparison makes that effort visible without pretending every hour returned will become a cash saving.

Find the assumption that can reverse the decision

Suppose a custom tool looks cheaper because its operating cost is mostly fixed, while the alternative charges for each user. The answer may depend on user growth, how much the business rules change and how much support the custom tool needs.

An estimate with one confident total hides those dependencies. A few plausible scenarios reveal them. If one uncertain assumption changes the preferred option, investigate that assumption before refining the rest of the arithmetic.

AI-assisted delivery may reduce implementation effort enough to change the comparison. It does not reduce every category by the same proportion. Migration, training, integration and ownership can remain substantial. Compare the complete commitments over the same period and workload, using current vendor terms when a purchase is actually being considered.

Include the work around the software

Organize the estimate into acquisition, operation, change and exit. Acquisition includes discovery, configuration, development, migration and training. Operation includes subscriptions, hosting, routine administration, maintenance and support. Change includes the likely improvements and adjustments the business will need. Exit includes export, migration, retraining and retiring the old arrangement.

Internal time belongs in the comparison even when it does not create a new invoice. A product that needs someone to reconcile records every week carries a continuing burden. Be clear whether you are estimating cash expenditure, staff capacity or both, so they are not accidentally counted twice.

The GOV.UK technology selection guidance is a useful external reference for considering choices beyond their initial implementation. Our emphasis is on the responsibilities the business will actually inherit.

Expose the uncertain assumptions

Use a range where the evidence is weak. Support demand, migration effort and the frequency of future changes may be uncertain. Show which assumption has the largest effect on the decision and investigate that one first.

For example, a custom tool might appear attractive because it avoids per-seat charges. That advantage could disappear if the planned feature set creates substantial maintenance work. A subscription product might appear expensive until its included support and administration features are accounted for. Neither conclusion can be reached from the headline price alone.

Connect cost to value

The cheapest option may be unable to deliver an essential outcome. The most capable option may provide features the business will never use. Compare each credible option’s cost with the value and constraints of the work it supports.

Keep a simpler alternative in the comparison, including a limited manual approach when appropriate. Build versus buy explores those alternatives, and measure value helps distinguish capacity returned from realized financial savings.

Update the estimate with experience

After launch, compare actual spending and effort with the assumptions. This can reveal an unexpectedly expensive integration, an unused subscription or a support need worth addressing through a small improvement.

The estimate becomes more valuable when it informs decisions throughout the system’s life. Keep it understandable enough that the owner can update it without commissioning another study.