Priority levels that people actually use correctly

Priority schemes fail in the same way everywhere: the top level becomes the default, because from where the requester sits their problem is the important one. The fix is not training. It is defining each level by what happens if the work waits, and putting that definition on the screen where the choice is made.

Four levels, defined by consequence

• •Now: somebody could be hurt, or part of the building cannot be used
• •Today: the business cannot do its work, or a legal requirement is being breached
• •This week: it will get worse or more expensive if it waits
• •Planned: none of the above

Four is the right number. Five levels invites a middle that means nothing. Three forces genuine problems into "planned" and they get lost.

Write the words on the form

Do not label the options P1 to P4, or High, Medium and Low. Those are internal names and everybody scales them differently. Put the sentence next to the option. A person choosing "somebody could be hurt" for a slow-draining sink knows they are stretching it.

Let triage change it, and say so

Somebody with context should be allowed to move the priority, and the requester should be told when that happens and why. The telling matters: a priority silently downgraded is how trust in the system dies.

Give each level a promise, and keep it

A priority with no response time attached is a label. Attach one, publish it, and measure against it. If you cannot meet the promise on the top level, the promise is wrong or the level is being used wrongly, and either way you have learned something.

Watch the mix

Count the proportion of work at each level, monthly. In a healthy building most work is planned and the top level is rare. If more than about one job in six is arriving as an emergency, you do not have a priority problem. You have a preventive maintenance problem wearing a priority costume.