What you measure, how often you look, and what that gap is costing you.
A structured review of your monitoring and reporting against a single question: are you looking at the right things, at intervals fast enough to catch them before they cost you money?
Most reporting cadences were not designed. They were inherited, from a time when a person had to do the looking, and the interval was set by what that person could sustain. The interval was never derived from the behaviour of the thing being measured.
That constraint has now moved. Almost nobody has gone back to re-examine what it was holding in place.
"Under-sampled data does not go quiet. It gives you a confident, stable, plausible picture of something that is not happening."
This is not an audit of whether your reporting is accurate. It almost certainly is. It is a review of whether the interval between one look and the next is short enough to catch the things that move fastest.
The same data, looked at too slowly. Nothing here is inaccurate. Every reading is correct. The trend they appear to form does not exist.
Each one is straightforward on its own. The finding comes from putting all three side by side, which is something almost no business has ever had reason to do.
Every measure that exists, and the reason each one exists. Some will turn out to be there because they were useful once. Some will turn out to be there because they are easy to produce.
Not the reporting calendar. The real interval: the moment when a person sits down, looks at a figure, forms a judgement about it, and decides whether to act. That is often a good deal less frequent than the reporting schedule suggests.
How quickly each underlying thing can move from healthy to costing you money. This is a question about your operation rather than about your data, and it is the one that is almost never asked.
In most businesses the result is the same in shape. A set of measures reviewed far more often than they can meaningfully change, consuming management attention for no return. Alongside them, a smaller set that can go wrong in days but is only examined monthly. That second set is where the money leaks.
A full inventory. Everything that is measured, how often it is reviewed, and how fast it moves.
A ranked list of detection gaps. Widest first, with an assessment of the exposure each one carries.
Recommended review intervals. For each measure, derived rather than assumed.
What can be corrected immediately. Using the systems you already have, at no further cost.
What would require development. Scoped and costed separately, so you know before you start what is and is not included.
The document stands on its own. It is written to be circulated internally and to support a budget decision without further explanation from me.
The Nyquist-Shannon sampling theorem holds that to observe something reliably you must sample it at least twice per cycle of the thing you are trying to detect. It has been settled since 1928.
Sample more slowly than that and you do not simply lose detail. You can be presented with a stable, plausible pattern that is not actually happening. The wagon wheel that appears to turn backwards in an old film is the everyday example. The wheel is not turning backwards. The camera is looking too slowly.
Classical sampling theory assumes a continuous band-limited signal, so applying it to business reporting is a motivated analogy rather than a theorem. It is a good analogy, and the arithmetic underneath it is real.
If you do not know your number, every reporting interval you run is a guess. It may well be a good guess. This is how you find out.
Start a Conversation