AI
← Back to library
Mental ModelsOpen4 min read

Your Dashboard Is Green But Your Gut Says No: The Blind Spot Finder

When every metric looks fine but something feels off, use this recipe to force the number that's hiding, before it becomes a fire.

The stuck moment. You run a 40-tech HVAC company. Revenue is up 12% over last summer, your booking rate is steady, review average holds at 4.7. Every number on the dashboard is green. But three of your best installers have gotten quietly short with dispatch, and last week a 10-year customer went with a competitor for a job you'd normally win in your sleep. Nothing you're measuring explains it, and that's exactly the problem. Your metrics are reporting on the business you had, not the one that's forming.

Why this happens: dashboards measure what already broke once and got instrumented. The next problem is, by definition, in the space you're not looking. Green metrics can coexist with a rotting core because you're averaging over the thing that's decaying.

The recipe (Blind Spot Finder, 5 moves):

  1. Name the unease in one sentence. "Things feel off" is too vague. Write the specific itch: "My best people are disengaging and I don't know why." Vague dread hides the signal.
  2. List what your dashboard actually measures. Revenue, bookings, reviews, close rate. Be honest and list only what you actually track today.
  3. Ask what each metric could hide. A steady close rate can hide which jobs you're winning (easy, low-margin) vs. losing (complex, high-margin). An average review score hides a shift in who is reviewing.
  4. Find the disconfirming number. For each suspicion, name the ONE metric you don't currently track that would prove or kill it. Margin-by-job-complexity. Repeat-customer win rate. Tech tenure vs. dispatch complaints.
  5. Pull that number first. Start with the one most likely to hurt. Look where you're afraid to look.

The AI move. Paste and run:

You are a skeptical operations analyst who assumes the dashboard is lying.

CONTEXT: I run a [business type, size]. My tracked metrics all look
healthy: [list your metrics + current values]. But my gut says something
is wrong. The specific unease is: [one sentence].

TASK: Find my blind spot. 
1. For each metric I named, tell me what a healthy-looking value could be
   HIDING (composition shifts, averaging effects, survivorship).
2. Generate the 5 most likely hidden problems that fit my unease but would
   NOT show up in my current metrics.
3. For each, name the ONE specific number or cut of data I should pull to
   confirm or kill it, and what result would mean trouble.

FORMAT: A ranked table: Suspected blind spot | Why my metrics miss it |
The number to pull | What bad looks like.

CONSTRAINTS: No generic advice. Assume I'm competent and the obvious stuff
is handled. Rank by (likelihood x damage). Lead with the one you'd check
first.

Take the top row and go pull that number today.

What it's worth: the gap between catching a margin leak or a talent exodus in week 3 versus quarter 3, usually five or six figures, and the two best people you didn't know you were losing.