BuzbyTimeRead the requirements guide

Getting people to record · 1.3

Reminders, nudges and escalation

Reminders, nudges and escalation. What decides it, what it costs, and what usually goes wrong. For another implementation reference, see Monitask's reference.

Managers ask for reminders. Reminders do not work.The usual request

A notification about an unfilled timesheet is a request to do the thing that was already unattractive, arriving at a moment chosen by software. It produces a small burst of low-quality entries and a steady erosion of goodwill.

What works instead is reducing the gap between doing the work and recording it, so that there is nothing to remind anybody about. For related guidance and industry context, see Trello.

Where a nudge is legitimate

  • At the end of the day, once, showing what was captured and inviting a correction.
  • Before a payroll or invoicing deadline, to the person, not to their manager.
  • When something looks wrong: an eleven-hour entry, a day with nothing, a category never used before.

Each of these gives the person something rather than asking for something, which is the distinction that decides how a notification is received.

Escalation to a manager

The commonest configuration and the one that converts a tool into a compliance regime. Once a manager is chasing entries, entries appear, and their accuracy is whatever is fastest to type.

If the data matters, protect its accuracy ahead of its completeness. A week with three honest days is more useful than a week with five invented ones.

Gamification

Streaks, badges and leaderboards attached to time recording. They lift completion in the first month and they attach a reward to the act of entering rather than to entering accurately.

What they reliably produce is a well-completed record that nobody trusts, and once a leaderboard exists it cannot be withdrawn without the withdrawal being read as a judgement.

Not offered

Leaderboards, streaks and any comparison between individuals are not offered, for the same reason across every page of this site.

What genuinely raises completion

  1. Entry that costs seconds.
  2. Something visibly useful coming back to the person.
  3. A manager who records their own time.
  4. Nobody being asked about an individual entry in the first month.

The third is underrated. A team whose manager does not record concludes accurately that this is for them rather than for the work.

When completion is bad for a good reason

Sometimes low completion means the categories do not fit, the work changed, or something is broken on a device. Asking why, once, without consequence, finds this. Escalating finds a number.

Notification fatigue is not a metaphor

A notification ignored twice is a notification that will be ignored permanently, and it takes the next one with it. This is well established in every domain that has studied alerting and there is no reason to expect this one to differ.

Which is an argument for one notification a day at most, and for removing any that does not produce action.

Measuring whether a nudge works

Look at whether entries appear within an hour of it, and whether that persists after the first fortnight. Most do not, and the ones that do not should be switched off rather than left running because somebody configured them.

The deadline that is real

Payroll and invoicing are genuine deadlines with consequences for other people, and a reminder tied to one is received differently from a reminder tied to a policy. Say what the consequence is: this goes to payroll on Tuesday.

What to do about the person who never records

Ask them, once, privately, why. The answer is usually specific: the tool does not work on their device, the categories do not fit their work, or they were burned by time data somewhere else.

Each has a response. Escalation has none, and it converts a solvable problem into a disciplinary one.

Deadlines that move

A payroll cut-off enforced one week and extended the next stops being a deadline. If the real cut-off is Wednesday, say Wednesday and hold it, and if there is genuinely flexibility, say that instead.

The digest that works

One message at the end of the week showing what was recorded, what it added up to, and where it went. It is a summary rather than a demand, people read it, and it produces corrections rather than resentment.

What to do first

Switch off every notification for a month and see what happens to completion. Most organisations find it barely moves, which settles an argument that otherwise runs for years.

The manager's digest

If anything goes to a manager, make it the aggregate rather than the individual gaps. A manager who receives a list of who has not filled in their timesheet will act on it, and that action is the thing this entry has been arguing against.

A closing thought

The measure of a good notification is whether removing it would be noticed. Most would not, and the ones that would are the two or three worth keeping.

Notifications people ask for

Worth distinguishing from the ones imposed. People frequently want a prompt at the end of the day, or a warning before a deadline, and a notification somebody chose is received entirely differently from one configured for them.

Make them optional by default and watch which ones people keep. That list is the answer to what the notification design should have been.

A rule of thumb

If a notification has been running for six months and nobody would complain about its removal, remove it. Applied once a year, this keeps the two or three that matter audible.

Also in getting people to record