Private BetaWe're currently in closed beta.Join the waitlist
All posts
ObservabilityJanuary 15, 2025

Exception-First Management: Why Your Dashboard Should Be Empty

The best operations surface only what needs attention. Everything else is noise.

Counterintuitive truth: The best dashboard isn't the one with the most metrics. It's the one that's usually empty—because everything is fine.

Most dashboards are museums of metrics. Rows of numbers, charts trending up and to the right, KPIs for every conceivable dimension. You open them, scroll, feel vaguely informed, and close them having taken no action.

This is backwards.


The Problem

TYPICAL DASHBOARD
┌────────────────────────────────────────────────────────┐
│ Revenue ████████████████████  Churn ███░░░░░░░░░░░░░░░ │
│ Tickets ██████░░░░░░░░░░░░░░  NPS   ███████████████░░░ │
│ Agents  █████████████░░░░░░░  Calls ████████████░░░░░░ │
│ ... 44 more metrics ...                                │
└────────────────────────────────────────────────────────┘
                          │
                          ▼
              "Uh... everything looks... fine?"

YOU: the filter, pattern-matcher, exception-detector

When everything is visible, nothing is visible. The cognitive load of processing all this information falls on you.

This doesn't scale. It doesn't even work well at small scale.


Exception-First Design

What if your dashboard was empty by default?

Traditional Dashboard
47 metrics displayed
You scan for problems
Maybe notice 1
Miss 2 more
Exception-First
0-3 items displayed
Each one needs action
Nothing healthy shown
Nothing missed

Not broken. Not missing data. Empty because everything is fine.

The only items that appear are exceptions—things that deviate from healthy:

  • Project went yellow → appears
  • Account untouched for 45 days → appears
  • Team member at 120% utilization → appears

When nothing needs attention, you see nothing. When something does, you see exactly that thing.


What This Requires

Exception-first design requires knowing what "healthy" looks like:

Metric Healthy Baseline Exception Trigger
Utilization 70-80% > 90% for 2+ weeks
Client contact Every 14 days > 30 days silent
Project velocity Within 10% of plan > 20% deviation
Response time < 4 hours > 24 hours

Once you have baselines, you can detect deviations. And deviations are the only things worth surfacing.


The Psychology of Empty

Empty inbox feeling: Everything is handled. Nothing needs you right now. Confidence without cognitive load.

This frees you to:

  • Do proactive work
  • Think strategically
  • Have the conversation you've been putting off

And when something does appear, you trust it. You know it matters. You act.


How to Build This

Start with one question for each thing you monitor:

"What would have to be true for me to not need to check this?"

If the answer is "metric is within range X," then:

  1. Stop checking it daily
  2. Set up an alert for outside range X
  3. Remove it from your regular review

Repeat for each metric. You'll find most don't need daily attention—they need attention when they break.


Key Takeaway

The goal isn't to be uninformed—it's to be informed efficiently. The best-run operations aren't the ones with the most dashboards. They're the ones where leaders spend the least time on operations because operations run themselves, escalating only when human judgment is needed.

Ready to see what your AI agents do?

Join the waitlist for early access.

Join Waitlist