## TL;DR
Founders read dashboards with six numbers, traffic-light colors, and a one-sentence takeaway. Not forty charts. Pick the numbers that signal health or trouble: volume trend, CSAT, first-contact resolution, backlog age, cost per ticket, and the top ticket driver. Auto-refresh it weekly, kill any chart nobody asks about, and always lead with what changed.

## The query

```text
building a support dashboard founders will read
```

## Use this when

- Dashboards exist but nobody opens them
- Reporting support health to leadership
- A founder asks "how is support doing" monthly
- You are drowning in charts with no readers

## Not for

- Real-time agent monitoring
- Deep-dive analytics
- Investor decks
- Day-to-day operational views

## Steps

### 1. Pick the six numbers

Volume trend, CSAT, first-contact resolution, oldest open ticket age, cost per ticket, and this month's top ticket driver. Six is the most a busy person absorbs. Every number must answer "are we okay" or "what changed." If a metric does neither, it does not belong.

Expected output: six agreed metrics, written down, with owners.

### 2. One chart per number, traffic-light colors

Each metric gets a single trend chart with green, yellow, and red zones marked. No dual axes, no stacked bars, no filters. A founder should grasp each chart in five seconds. If a chart needs explaining, it failed.

Expected output: six dead-simple charts a non-support person reads instantly.

### 3. Lead with the one-sentence takeaway

Top of the dashboard, one line: "CSAT dipped on billing tickets after the pricing change; fix ships Thursday." The takeaway is what gets read. The charts are the evidence for anyone who wants to verify.

Expected output: a current, honest headline updated with every refresh.

### 4. Auto-refresh weekly, same day

Stale dashboards train people to ignore dashboards. Automate the data pull so the numbers are fresh every Monday without human effort. A dashboard that needs manual updating will be abandoned by March.

Expected output: a dashboard that is never more than a week old, with zero manual work.

### 5. Kill what nobody asks about

Every quarter, check which charts get discussed and which get scrolled past. Delete the ignored ones. A dashboard that shrinks toward what matters stays read; one that accumulates charts becomes wallpaper.

Expected output: a dashboard that gets smaller and more read over time.

## Variant phrasings

### customer support KPI dashboard

Steps 1 and 2. KPI dashboard is the corporate name for the same six numbers.

### executive support metrics report

Steps 1, 3, and 4. Executives want the takeaway and the freshness; the charts are backup.

### support health dashboard template

The full five steps produce the template. Health means the traffic lights, not the raw numbers.

## Why it happens

Support people build dashboards for themselves: detailed, filterable, comprehensive. Then they show them to founders, who see noise and stop opening them. The founder's question is never "show me everything"; it is "are we okay, and what changed." Six numbers with colors and a headline answer that question. Everything else is for the support team, and it belongs on a different screen.

## Edge cases

- The founder wants more detail: link the six charts to the deep dashboards. The top screen stays simple; depth is one click away.
- Numbers disagree with the founder's anecdotes: trust the numbers, but investigate the anecdote. One loud complaint can reveal a real segment problem the averages hide.
- Early stage with tiny volume: percentages swing wildly. Show raw counts alongside rates until volume stabilizes.
- Multiple products or regions: one dashboard per major segment, or the blends will lie. Do not mix enterprise and self-serve.
- The takeaway writer leaves: make the headline part of the weekly metrics review ritual, not one person's talent.

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_T_W6wJtVNGSkLbZ-s_iZlQ