Some of our weekly metrics fluctuate wildly from week to week, causing our leadership team to panic and try to solve non-existent problems during our Level 10 Meeting. How do we distinguish between normal operational noise and a genuine trend that requires us to IDS the issue?

Question

Tyler Smith · Accepted Answer

Scorecard panic is a common symptom of tracking the wrong time horizons. If a single bad week sends your team into a tailspin, you are missing the forest for the trees. You must establish rules that separate normal weekly variance from a systemic issue that requires a deep-dive IDS.

To differentiate between noise and signal, implement these three guidelines:

- Look at the thirteen-week trend line: Never analyze a single red week in isolation. Look back at the last three to four weeks. If a number is red once but has been green for the previous twelve, it is likely noise. If it is red three weeks in a row, it is a trend.
- Define acceptable variance thresholds: Establish clear upper and lower limits for your metrics. If a number drops but remains within your acceptable variance band, do not drop it to IDS. Only drop it if it breaches the floor.
- Track trailing averages on the scorecard: For highly volatile metrics, track a trailing four-week average alongside the weekly number to smooth out the spikes.

Your Level 10 Meeting is for solving systemic operational issues, not reacting to daily randomness. Teach your team to respect the trend lines, and only IDS a missed number when the data proves there is a broken process underneath.