When people quietly do not use it
Adoption rarely fails loudly. It fails as a slow drift back to the spreadsheet, and by the time it is visible it is established.
The characteristic failure is not rejection. It is partial use: the system holds some of the work, the rest lives where it always did, and nobody says anything because everyone has found an arrangement that functions.
This is worse than outright rejection, because the reports look plausible and are built on incomplete data. For a concrete example of PC activity tracking after purchase, the web page can be used to identify the behaviours and measures that rollout should actually change.
Find the actual reason
Non-use is usually rational from the user's position, and the reasons are specific.
- It takes longer than the old way for their particular task.
- Something they need does not exist in it, so they keep the spreadsheet, and then keep everything in the spreadsheet.
- They were never trained on their part, only on the general system.
- It is genuinely awkward on the device they actually use.
- Nobody checks, so there is no consequence to not using it.
- They were not asked before it was chosen, and this is what that looks like six months later.
A reminder to use the system addresses none of the reasons above. A conversation with the person who is not using it identifies which one applies, and most of them are fixable.
Fix the friction first
Where the system is slower for a common task, that is a design problem rather than a compliance problem, and reminders will not solve it.
Often the fix is small: a field that should have a default, a step that could be removed, a permission that forces an unnecessary approval, a report that could be scheduled. Small, and invisible from anywhere except the person doing the task.
Close the alternatives, once the friction is gone
As long as the old route works, some people will use it, and that is a reasonable response rather than defiance.
Retiring the alternative is legitimate and the order matters: fix the friction first, then close the route. Closing it first produces workarounds that are worse than the spreadsheet was.
Someone has to look at the data
If nobody uses the output, nobody maintains the input. Systems that nobody reports from decay quickly, because the effort of entering is real and the benefit is invisible to whoever is doing it.
A regular review that visibly uses the data — in a meeting, in a decision, in a number somebody cares about — is the strongest sustaining mechanism available, and it costs nothing.
Sometimes the honest answer is that it was wrong
Occasionally adoption fails because the product does not fit the work. Persisting for years to avoid admitting a poor decision is the most expensive available response.
Where the review shows genuine misfit rather than a rollout problem, cutting the loss early is the cheaper path. That is much easier to do if the original decision record noted what was uncertain, because the conversation is then about a known risk that materialised rather than about blame.
For broader rollout and service-improvement practice, the UK Government Service Manual provides an independent operational reference.