Running a trial that tells you something
Trials are run in the quietest month, by the most capable person, on twenty tidy sample assets. That proves nothing about February.
The standard maintenance software trial is conducted by whoever is most comfortable with software, on a small set of clean sample assets, during a quiet period. Everything works, the purchase is made, and the product meets the real estate three months later. For a separate product-side perspective on recognising artificial activity signals, this analysis provides additional detail.
Running a trial properly means deliberately reproducing the conditions the system will actually operate in.
Load real data, including the messy parts
Sample data is clean. A real asset register contains duplicates, assets with no serial number, three different naming conventions from three different eras, a building that was renamed, and locations recorded as 'plant room'.
Loading the real thing is the only way to find where the product's data model conflicts with yours, and it also gives an honest measure of the migration effort — which is regularly underestimated several times over. For a wider operational and compliance reference, consult CISA operational technology guidance.
Test the awkward scenarios
- A statutory task that must not move, alongside a task that can float.
- An asset replaced with a different model, keeping the history.
- A job attended, made safe, awaiting a part for three weeks, then completed.
- A callout at two in the morning at a premium rate.
- A certificate attached to an asset and retrieved two years later.
- A job recorded offline in a basement and synced two hours later.
Each of these is ordinary rather than exotic, and each has caused an organisation to discover a limitation after signing.
Let a technician drive
The evaluation is usually run by a manager, which produces a verdict about the manager's experience. Have a technician who did not choose the product complete the core tasks unaided: find the job, see the asset history, record what was found, close it, on their own phone, in a plant room.
If they cannot, adoption will be a two-year argument regardless of how good the reporting is.
Not the newest phone on office wifi. The oldest one, in the basement, with gloves on. That is the environment the data will be created in.
Test support before you need it
Raise a genuine question during the trial and time the response. Note whether the answer comes from someone who understands maintenance or from a script.
Ask specifically what happens when a statutory frequency changes, who updates any regulatory content the product ships with, and whether that is included in the subscription.
Do the export, for real
Establish the notice period, what happens to your data at the end of the contract, and how long access continues. Then run a full export during the trial and open the files.
This takes about half an hour and is the single most informative test available, because it is the only one the vendor has no incentive to make easy. Check whether attachments — certificates, photographs — come out with the records or are left behind as filenames.