A data problem is not a signal problem
Most organisations answer a signal problem with more data. It is the most expensive mistake in modern management.
Two problems that look identical
A data problem means you do not have enough information. The answer is to go and get more.
A signal problem means you already have too much, and none of it is telling you what to do. The answer is the opposite: remove, filter, distil.
These two are confused constantly, and the confusion is expensive.
The wrong cure
Most businesses respond to a signal problem as if it were a data problem. They buy another tool. They add another report. They track ten more metrics. They pour more water into a glass that is already overflowing, and then wonder why they feel like they are drowning.
The answer to too much noise is never more noise.
Leaders sit in front of dashboards with forty metrics and still make the real decisions on instinct, because nothing on the screen tells them which number actually matters.
How to tell which one you have
Ask three questions about the information in front of you:
- Decision relevance. If knowing it would not change what you do, it is noise — however interesting it is.
- Time sensitivity. A true signal is tied to a moment. Most information is history dressed up as insight.
- Actionability. If no one owns it and no action follows, it does not belong in front of a decision maker.
What survives all three is signal. Everything else, no matter how much effort produced it, is noise wearing a suit.
The move
Before you commission another report, name the decision it serves. If you cannot name one, you do not have a data problem. You have a signal problem, and the cure is subtraction.
From Signal over Noise: The Playbook.