How to Build a Product Health Scorecard in Notion
Product health is not a single number — it is a composite view across engagement, quality, growth, and customer satisfaction. Most PMs track these metrics in four different tools and compile them into a slide deck every week. By the time the deck is ready, the data is two days stale and the context behind the numbers is lost.
A Notion-based product health scorecard centralizes these metrics in a living document that the whole team can see and contribute to. This guide walks through building a scorecard that auto-updates where possible, uses color-coded status indicators for at-a-glance readability, and includes the narrative context that raw numbers lack.
Step-by-step guide
Define the four health dimensions and their metrics
Create a Notion page and start with a table that defines your health dimensions: Engagement (DAU/MAU, session duration, feature adoption), Quality (error rate, p95 latency, crash rate), Growth (signups, activation rate, MRR), and Satisfaction (NPS, CSAT, support ticket volume). For each metric, define the source, the healthy range, and the warning threshold.
- List 2-3 metrics per dimension (eight to twelve total is manageable for weekly review)
- Define green/yellow/red thresholds for each metric
- Document the data source and owner responsible for updating each one
Create the scorecard database
Build an inline database with columns: Metric Name (title), Dimension (select: Engagement/Quality/Growth/Satisfaction), Current Value (number), Target (number), Status (select: Green/Yellow/Red), Trend (select: Up/Flat/Down), Last Updated (date), and Notes (text). Each row is one metric. The Status column gives you the at-a-glance view; the Notes column provides the narrative.
- Create an inline database on your scorecard page
- Add all properties with appropriate types
- Pre-populate with your defined metrics and their current values
Build views for different audiences
Create three views of the database: an 'Executive Summary' board view grouped by Status (so red items are immediately visible), a 'By Dimension' table view grouped by the Dimension property, and a 'Trends' gallery view showing only metrics where Trend is 'Down'. The executive summary is for leadership; the dimension view is for the PM team; the trends view is for triage.
- Create a Board view filtered to show only Yellow and Red metrics for quick triage
- Create a Table view grouped by Dimension for the full picture
- Create a Gallery view with a filter on Trend = Down for degradation alerts
Add sparkline charts or embedded analytics
For each metric, embed a trend chart from your analytics tool (Amplitude, Mixpanel, or Google Analytics all support embed URLs) or use Notion's chart feature to visualize the trend. A number without a trend is hard to interpret — is 45% DAU/MAU good? Only the trend tells you whether things are improving or degrading.
- For each metric row, add an embed block with the chart URL from your analytics tool
- If embeds are not available, use Notion's chart block with manual data entry
- Update charts weekly as part of the scorecard update process
Set up the weekly update cadence
Assign a 'Scorecard Owner' who updates all metrics every Monday morning before the leadership sync. Create a Notion template for weekly snapshots: a page titled 'Week of [Date] — Product Health' that captures this week's scorecard state plus a narrative section explaining any red or degrading metrics. This creates a searchable history of product health over time.
- Create a 'Weekly Snapshots' database linked to the scorecard
- Build a template with auto-populated date and sections for each dimension
- Set a recurring Slack reminder for the scorecard owner to update by Monday 9am
Connect the scorecard to action items
Add a 'Response' relation property that links red and yellow metrics to the Notion task or Linear issue that addresses the degradation. This closes the loop: a metric turns red, someone creates a task to investigate, and the task is linked to the metric so leadership can see 'this is red AND someone is working on it' vs 'this is red and nobody has noticed.'
- Create a relation to your team's task database or project tracker
- Make it a policy that every red metric must have a linked response within one business day
- Add a 'Has Response' formula column that shows a checkmark when the relation is not empty
Common mistakes
Including too many metrics and losing signal in noise
A scorecard with thirty metrics is not a scorecard — it is a data dump. Keep it to eight to twelve metrics that genuinely reflect product health. If a metric has been green for six months and nobody looks at it, remove it and add something more informative.
Setting thresholds based on aspirations rather than baselines
If your DAU/MAU is currently 25% and you set the green threshold at 40%, the metric will be permanently red and the team will stop paying attention. Set thresholds based on your current baseline plus realistic improvement targets. Recalibrate quarterly.
Updating numbers without narrative context
A metric that drops from green to yellow without an explanation generates anxiety, not action. Always update the Notes column when a metric changes status. 'DAU dropped 15% due to holiday week — expected to recover next week' is infinitely more useful than a yellow dot.
Tips
Use Notion's conditional formatting (formula + color) to make the Status column auto-calculate based on Current Value vs Target — no manual color-coding needed.
Create a 'Historical Values' database that stores weekly metric snapshots so you can plot trends over months, not just see the current week.
Pin the scorecard to your team's Slack channel topic so anyone can find it without asking.
Review and prune the scorecard quarterly — remove metrics that have been consistently green for three months and add new ones that reflect current product challenges.
How Vantage helps
Vantage's analytics integration surfaces product health data directly within your PRD generation context. When you start a new project targeting a degrading metric, Vantage already knows the historical trend, current value, and related past initiatives, so the generated PRD is grounded in the real health picture rather than a vague sense that 'retention needs work.'