Everyone is using the system. Almost nobody is using it while the work is happening.
15/09/2026
Adoption reporting typically measures completeness: what proportion of the work was recorded in the system. The more revealing measure is timing. When did the entry happen, relative to the event it describes? A system used correctly but always retrospectively is not carrying the work. It is functioning as a reporting layer over an unchanged process,
Where does your adoption figure come from?
15/09/2026
Somebody reports an adoption figure to your steering committee. Does anyone in the room know what its denominator is? Every adoption percentage carries one, and it is usually chosen by accident. Does the population include everyone with an account, everyone holding the role, or only those who received eligible work in the period? Are people
The implementation failures that cost the most are the ones you will not find for two years.
15/09/2026
They are not dramatic. They are quiet, competent adaptations that leave every formal indicator intact. Which is why acceptance risk needs a third dimension beyond likelihood and impact. If this happens, how would the organisation find out, and when? A risk that announces itself is manageable almost regardless of its likelihood. A system nobody can
“Users may resist the new system.” If that sentence is in your risk register, you do not have a risk. You have an anxiety.
15/09/2026
It contains no cause, no specific behaviour, no consequence and no group. It cannot be assessed for likelihood, cannot be assigned to anybody meaningfully, and cannot be closed except by assertion. A usable acceptance risk states three things. The cause: something identifiable about the design, the process, the environment or the distribution of burden. The
The project was recorded as a success. The benefit never arrived. Both statements are true.
15/09/2026
An implementation produces outcomes at four levels, and only one of them is normally contracted. The technical outcome: the system exists and functions within agreed parameters. The operational outcome: people can use it to perform their work, including the interruptions, exceptions and time pressure the process documentation left out. The behavioural outcome: they do. The
The adoption report says 91 per cent. So why is nothing different?
15/09/2026
An implementation is delivered, the project closes, and the reports look acceptable. Two years later the benefit has not appeared and nobody can say exactly why. Part of the answer is that non-adoption is rarely refusal. It takes four recognisable forms, and only one of them is visible in a standard adoption report. Ceremonial use.
The software works. The project closed on schedule. So why is everyone still using the spreadsheet?
14/09/2026
Most IT implementations that fail do not fail loudly. The software is delivered. The tests pass. The system goes live more or less on schedule, and the steering committee records a successful implementation and closes the project. Eighteen months later, the spreadsheet is back. The reports cannot be relied on. Somebody is proposing a project