Automated employee reports: forget Excel

Automating a report that nobody acts on produces the same inaction, faster. Before building anything it is worth writing one sentence per report: who reads it, what decision it supports, and what they would do differently if the number moved.

Reports that survive that sentence are usually a small minority, and that is the useful finding. Most recurring reports exist because someone once asked for them, and the person who asked has moved on. Retiring them frees the attention that the remaining ones need.

A recurring report answers a question on a schedule. An alert answers it when the answer changes. Most of what firms send weekly should be an alert, and most of what they send as alerts should be a report — the inversion is where notification fatigue comes from.

The rule that works: if the recipient will act only when a threshold is crossed, send an alert with the threshold in it. If they need context and trend to make a judgement, send a report on a rhythm. Sending both for the same number guarantees neither is read.

An automated report loses credibility the first time someone finds a figure they cannot reconcile, and it never fully regains it. The causes are mundane: a definition changed without announcement, a data source that stopped updating, a filter that silently excludes a team.

Two habits prevent almost all of it. Every report states its definitions and the date its data was last refreshed, and every definition change is announced before it appears. It costs a line of text and it is the difference between a number people act on and a number they check by hand first.