A team dashboard people actually open
Stop the 'where's the number?' Slack pings — give the team one screen they trust.
What you'll have at the end
A single dashboard your team opens on purpose, not because you nagged them. The numbers match the source of truth, the page loads in under two seconds, and every chart answers a question someone actually asks on a Monday.
01
Who this is for
- Founders and operators tired of pasting the same numbers into Slack every Monday
- Team leads who keep getting 'what's our X right now?' DMs
- Builders who already shipped a small internal tool and are ready to layer real reporting on top
02
How to frame the idea
A dashboard is not 'every chart we could draw.' It's the answer to the three questions the team asks most often, on one screen, refreshed often enough to trust. Write those three questions down first. Build a chart for each. Resist every urge to add a fourth until someone asks for it twice.
03
What people actually build
Weekly ops snapshot
Revenue this week vs last, active customers, open tickets, top blocker. One screen, refreshed nightly, replaces three Slack threads.
Sales pipeline at-a-glance
Deals by stage, value, owner, and slip risk. Whatever's in your CRM, summarized so nobody has to log into the CRM to answer 'how are we tracking?'
Product health board
Daily active users, signups, errors per 1k requests, latency. The numbers engineering and growth both stare at — but neither has to build twice.
04
Tool choices, honestly
Lovable + Lovable Cloud
You want the dashboard, the auth, and the data pipes in one project. Best when the numbers come from data already in your app.
Lovable + read-replica / warehouse
Numbers live in production tables you don't want a dashboard hammering. Point at a replica or a small warehouse (BigQuery, Snowflake, Postgres replica) and query that.
Metabase / Looker Studio
Honestly — if the audience is internal-only and tolerates a generic UI, an off-the-shelf BI tool will get you 80% there in an afternoon. Build custom when the dashboard is part of the product or the brand.
05
Prompts you can lift
Pick the three questions
Help me pick the three questions a team dashboard should answer. Here's the team and what they do: <one paragraph>. Here's what people DM each other about most: <three things>. Propose three questions, the chart type that best answers each, the data source, and the refresh cadence. Reject any chart that doesn't directly answer one of the three questions.
Define the metrics precisely
For each metric on the dashboard, write a one-sentence definition I could paste into a tooltip: what it counts, what it excludes, the time window, and the source table. Flag any metric where two teammates would reasonably define it differently — those need a decision before I build the chart.
Build the read path safely
Build the dashboard read queries against <data source>. Cache results so a page load doesn't run the queries from scratch. Show me the cache TTL per chart, the invalidation strategy, and a way for a teammate to force-refresh when they don't trust the number.
Lock down who sees what
Add auth. Anyone on the team can see the top-level numbers. Only finance can see revenue detail. Only engineering can see error rates by service. Show me the exact role rules and how I'd test them by signing in as each role.
Save and reuse these prompts in PromptlyDo™ with your favorite AI.
- Install the PromptlyDo™ browser extension
- Sign in or create a free account
- Right-click any prompt above and save it to PromptlyDo™
06
What tends to break
- Building 12 charts because you could, instead of 3 charts because someone asked. The 12-chart dashboard gets opened twice, then never.
- Numbers that disagree with the source of truth. The CFO checks the dashboard against the CRM, finds a $2k gap, and never trusts the dashboard again.
- Slow queries that run on every page load. The dashboard becomes the reason your production database is on fire.
- Charts with no definition. Two teammates argue about 'active users' for an hour because nobody wrote down what it means.
- No 'as of' timestamp. People stare at last Tuesday's snapshot thinking it's live.
- Permissions as an afterthought — the intern can pull up the revenue chart on day one.
07
What AI forgot to ask you
- Live or daily? 'Real-time' is expensive and almost never what the team actually needs. Nightly is usually enough.
- What's the source of truth for each number, and what happens when it disagrees with another source?
- Who's allowed to see revenue, headcount, churn, customer names? Default-open is a leak waiting to happen.
- What does the dashboard show on day one for a brand-new team or empty period? Zeros, dashes, or 'not enough data yet'?
- Where does the dashboard live — its own URL, embedded in the main app, or both? That changes auth and branding.
08
Before real users see it
- Every number on the dashboard matches the source of truth to within rounding — I spot-checked three of them by hand.
- Every chart has a one-line definition visible on hover or click.
- Every chart shows an 'as of' timestamp.
- A teammate without admin access cannot see the metrics they shouldn't — I tested by logging in as each role.
- Page loads in under two seconds on the slowest expected query, with cache warm.
- There's a 'this number looks wrong' link or button that opens a way to flag it without DMing me at 11pm.
09
Questions to sit with
- 1.What's the one question this dashboard exists to answer — and if it could only show one chart, which would it be?
- 2.Who is supposed to take action when a number turns red, and have I told them that out loud?
- 3.Am I building a dashboard, or am I building a status page the team treats as a dashboard? Those are different products.
- 4.If I removed the bottom half of the dashboard tomorrow, would anyone notice — and what does that tell me?
Ready to app it?
Take this path into your tool of choice — and when you finish (or get stuck), share what you learned so the next builder doesn't reinvent it.