What goes on a product team dashboard?
A product team dashboard follows users through the whole journey — signup, activation, adoption, retention — so the team can see where people find value and where they slip away. This page is the hub: a live example you can click around, the dashboards a product team typically builds, and the metrics worth putting on them.
For: product managers, product analysts, growth, and design. What’s inside: a live public Metabase dashboard, six dashboards a product team builds, ten metrics that belong on them, and a five-step build order.
Explore a live product dashboard
This is a real Metabase, not a screenshot. Click into any card to see the question behind it, change the filters, hover the series, or drill from a bar down to the underlying rows — it behaves exactly the way your own dashboard would.
Which dashboards does a product team build?
Most product teams end up with one hub covering the journey and a few dashboards underneath it for the parts that need depth. Build the retention view early — it is the number that decides whether anything else you do compounds — and add the qualitative pages once you have enough feedback volume for themes to be visible.
Which metrics belong on a product team dashboard?
Ten that earn their place on the hub. Read them in journey order — the top half tells you whether people arrive and get going, the bottom half whether they stay and are worth more over time.
- Signup conversion rate — the share of visitors who create an account. Watch it by landing page and source; a blended number mostly reports where your traffic came from that week.
- Signups by channel — new accounts split by source or campaign. The point is not volume but quality: two channels with identical signup counts routinely activate at half the rate of one another.
- Activation rate — the share of new users who reach their first real outcome. This is the single card most worth arguing over, because it gates every number below it.
- Time to value — how long activation actually takes. A flat activation rate with a lengthening time to value is an onboarding problem that has not shown up in the headline yet.
- Trial-to-paid conversion rate — how many trials become paying customers. Cohort it by trial start date rather than conversion date, or a good month of signups will look like a bad month of conversions.
- Feature adoption rate — what share of eligible accounts use each key feature. Pair breadth with depth; a feature 60% of accounts tried once is a different result from one 20% use weekly.
- Cohort retention — the share of each signup cohort still active week by week, on the cohort retention grid. If the curve flattens, you have a product people keep; if it keeps sliding, nothing upstream will save it.
- Churn rate — customers or revenue lost each period. Split voluntary from involuntary before you draw any product conclusion; failed cards are a billing fix, not a roadmap item.
- Net revenue retention — expansion minus contraction and churn across existing customers. It is the one product metric your board already tracks, and above 100% it means the product grows without new logos.
- Net promoter score — how likely customers are to recommend you, sitting next to CSAT as the slow-moving sentiment counterweight to a page of behavioural counts.
How do you build it?
- Get product events into the warehouse alongside billing and CRM data. Metabase queries them where they sit, but activation and NDR need events joined to subscriptions, so the two have to live somewhere they can meet.
- Define user, account, active, and activated once, in Metabase models that every card reads from. Product teams argue about definitions more than any other function, and a shared model is what turns those arguments into a one-time decision.
- Build the journey spine first — signups, activation, weekly active accounts, retention — as one column of cards in order. Anything that is not a stage or a rate between two stages goes below the fold.
- Add the value layer: feature adoption for the two or three features that matter this quarter, plus churn and net revenue retention so the page ends on whether the product is worth more over time.
- Add filters for date range, plan, and segment, drop in one survey card, then schedule the dashboard to the product channel before each planning cycle so the roadmap conversation starts from the same numbers.
Related
Integrations
Other team dashboards
FAQ
How is a product team dashboard different from a feature dashboard?
Scope. A feature dashboard — feature adoption and engagement, a flag rollout — answers "did this thing land", and it is usually read hard for six weeks and then archived. A product team dashboard follows the journey end to end: signup, activation, adoption, retention, and the revenue that survives it. It is the page you open every week for years, so it should hold the shape of the funnel rather than the detail of whatever shipped last sprint.
Who owns a product team dashboard?
A product analyst or the PM who cares most builds it; the head of product owns what stays on it. The predictable failure is one card per squad, added the week that squad wanted visibility and never removed. Keep the hub to the journey stages every squad shares, give squads their own dashboards for their own surfaces, and prune the hub each quarter — the test is whether anyone would notice the card missing.
How often should it refresh?
Daily for the acquisition and activation half, weekly or monthly for retention and revenue. Product events land in near real time, so a daily refresh is cheap, but reading cohort retention daily is noise — a 30-day cohort is not comparable until 30 days have passed. Split the difference by putting the fast-moving cards at the top on a daily schedule and setting the retention and NDR cards to a weekly grain with a clear as-of date on the card.
Why don't the numbers match Amplitude, Mixpanel, or PostHog?
Because those tools count events their own way and drop what they sample. Session boundaries, bot filtering, and identity stitching all differ; a product analytics tool also usually starts counting at first tracked event, while your warehouse starts at row creation. Land the raw events in the warehouse, define signup, activation, and active user once as Metabase models, and treat the vendor UI as an exploration surface rather than a source of truth. The Amplitude, Mixpanel, and PostHog guides cover getting each export landed.
How do you define activation?
Pick the earliest action that predicts retention, then never change it quietly. For most products it is a small sequence rather than a single click — created a project, invited someone, saw a result — and the honest way to choose it is to look at what your retained cohorts did in week one that the churned cohorts did not. Write the definition into one model, version it, and if you do change it, rebuild history on the new definition so the activation rate line does not step for a reason nobody remembers.
Should qualitative feedback be on the same dashboard?
Two or three cards of it, yes. A page of pure behavioural numbers tells you what happened and never why, and the fastest bridge is an NPS or CSAT trend next to a count of open feature requests by theme. Keep the volume low — the full qualitative picture belongs on voice of customer and feature request pipeline — but having one survey card beside the funnel is what stops a retention dip turning into a week of speculation.