Systems Outlive People
Every system becomes a fossil record of the decisions that shaped it. Long after the people who built it have moved on, the system quietly preserves their assumptions, compromises, and priorities.
Read →Most reporting problems don’t begin with data quality.
They begin at the last mile, where information is supposed to influence a decision.
It’s common to see technically sound dashboards that still fail to change behaviour. Not because the numbers are wrong, but because the report doesn’t align with how decisions are actually made.
Where Things Break Down
Last-mile failures usually look like this:
- Metrics without clear owners
- Charts without decisions attached
- Data that updates regularly but isn’t acted on
- Reports built around availability, not usefulness
The output exists.
But the purpose is vague.
Reporting Isn’t the Same as Insight
Insight requires context:
- What decision is this supporting?
- What action should follow?
- What happens if the number changes?
Without that framing, reporting becomes passive.
People look at it, nod, and move on.
Designing for Decisions
The most effective reporting starts with questions, not data:
- What decision is someone trying to make?
- What would change their mind?
- What do they need to trust before acting?
When those answers are clear, the data requirements often become simpler, not more complex.
Clarity reduces noise.
Reporting That Works
Good reporting doesn’t get talked about much.
It gets used.
When decisions happen without debate over the numbers, that’s usually a sign the last mile has been designed deliberately.
Because reporting doesn’t succeed when it looks impressive.
It succeeds when it changes behaviour.
Related posts
Systems Need Owners, Not Just Users
Systems degrade when everyone uses them but no one owns them. Without clear accountability, definitions drift, exceptions multiply, and trust quietly erodes.
Read →Automation Doesn’t Start With Tools
Automation fails when it begins with procurement instead of process clarity. Before software, there must be visibility, ownership, and agreement on what should - and shouldn’t - be automated.
Read →