Metrics, OKRs & Product Analytics
InsightFlow's Workflow Builder Sprint 1 shipped. The trigger configuration UI is in design testing. Stakeholders have been updated. Now comes the question that distinguishes a PM who is busy from a PM who is effective: did any of this actually move the metrics that matter?
You have data from the past month. New signups are up. But the North Star metric — weekly active teams creating or editing dashboards — is flat. Activation rate (the percentage of new signups who reach their first dashboard) dropped slightly. Enterprise retention is strong, but free tier churn is above the benchmark for Series B companies in this space. There is something worth investigating here, and a monthly metrics review is the structure for finding it.
The /metrics-review command from the official product-management plugin turns raw metric data into a structured analysis: a scorecard with comparisons, trend analysis, bright spots, areas of concern, and recommended actions. But the command's value is proportional to how well you have defined your metrics hierarchy — because a review against the wrong metrics is not just useless, it is actively misleading.
This lesson builds the hierarchy first. Then runs the review.
The Product Metrics Hierarchy
A well-designed product metrics system has three levels:
North Star Metric (1)
│
▼
L1 Health Indicators (5-7)
Acquisition | Activation | Engagement | Retention | Monetisation | Satisfaction
│
▼
L2 Diagnostic Metrics (as many as needed)
Feature-level usage | Funnel conversion steps | Segment breakdowns
North Star Metric
The single metric that best captures the core value your product delivers. It should be:
| Criterion | Description | InsightFlow Test |
|---|---|---|
| Value-aligned | Moves when users get more value | Yes — users get value when they create/edit dashboards |
| Leading | Predicts long-term retention and revenue | Yes — dashboard creation predicts ongoing engagement |
| Actionable | The product team can influence it | Yes — onboarding, templates, and performance all affect it |
| Understandable | Everyone knows what it means and why | Yes — "active teams building dashboards" is clear |
InsightFlow's North Star: Weekly active teams creating or editing dashboards
Why not MAU? Monthly Active Users always grows with marketing spend. It does not tell you whether users are getting value — only that they logged in. A team could grow MAU while the core product experience deteriorates. The North Star should move when users genuinely benefit from the product.
Why not revenue? Revenue is a lagging indicator — it measures what happened 3-6 months ago in SaaS, not what is happening now. By the time revenue trends appear, the underlying product health signals have been visible for weeks.
L1 Health Indicators
The five to seven metrics that together paint a complete picture of product health. They map to the user lifecycle:
| L1 Indicator | What It Asks | InsightFlow Metric |
|---|---|---|
| Acquisition | Are new users finding the product? | New trial signups per week |
| Activation | Are new users reaching the value moment? | % of signups who create their first dashboard within 7 days |
| Engagement | Are active users getting value regularly? | North Star — WAT creating/editing dashboards |
| Retention | Are users coming back? | D30 retention: % of signups still active at 30 days |
| Monetisation | Is value translating to revenue? | Free-to-Pro conversion rate; MRR growth |
| Satisfaction | How do users feel about the product? | NPS; export/dashboard-specific CSAT |
Activation rate — the percentage of new signups who reach the first meaningful value moment — is typically the highest-leverage metric a PM can improve. Low activation means users are signing up, trying the product, and leaving before they get value. Fixing activation has a compound effect: it improves retention, LTV, and NPS simultaneously.
L2 Diagnostic Metrics
L2 metrics answer "why is an L1 metric moving?" They are not for the weekly dashboard — they are for investigation.
If activation rate drops (L1), L2 diagnostics might include:
- Funnel conversion at each onboarding step
- Time-to-first-dashboard for this week's cohort vs. prior cohorts
- Feature adoption: are new users finding the template gallery?
- Device and browser breakdown (is there a regression on mobile?)
L2 metrics are selected based on the question. You should not have a standing L2 dashboard with 50 metrics. You should have the analytical tools to investigate when an L1 signal changes.
OKRs: Writing Key Results That Actually Measure Outcomes
OKRs (Objectives and Key Results) are the goal-setting system that connects your product priorities to measurable outcomes. They fail in one consistent way: teams write Key Results that measure what they ship, not what users do.
| OKR Component | Definition | Test |
|---|---|---|
| Objective | Qualitative, aspirational. What does success feel like? | Is it inspiring? Does it capture a direction? |
| Key Result | Quantitative, measurable. How do you know you achieved the Objective? | Is it outcome-based? Can users benefit without this changing? |
The output KR trap:
| Output KR (Wrong) | Outcome KR (Right) |
|---|---|
| "Ship 3 new features" | "Increase activation rate from 35% to 50%" |
| "Complete 10 customer interviews" | "Reduce D30 churn from 18% to 12%" |
| "Rebuild the onboarding flow" | "Improve time-to-first-dashboard from 4 days to 2 days" |
| "Launch Workflow Builder beta" | "50 beta users complete their first automated workflow" |
The test for an output KR: could your team hit this KR without a single user benefiting? If yes, it is the wrong thing to measure.
Well-formed OKR example for InsightFlow Q2:
Objective: Make InsightFlow the default choice for data analysts
who need dashboards without SQL
Key Results:
KR1: Increase North Star (WAT creating/editing dashboards) from
520 to 750 teams per week
KR2: Improve activation rate (% of signups who create first
dashboard within 7 days) from 35% to 50%
KR3: Reduce free-tier D30 churn from 18% to 12%
KR4: Achieve NPS ≥ 45 (up from current 38)
OKR Quality Rules
- 70% completion is the target for stretch OKRs — if you are confident you will hit 100%, they are not ambitious enough
- 2-3 Objectives maximum per quarter — more than that and nothing is truly prioritised
- Key Results should be uncomfortable — they should require genuine product improvement, not just shipping
- Review OKRs at mid-quarter — if a KR is clearly off track, adjust effort allocation or flag it as a risk
The /metrics-review Command
The /metrics-review command generates a structured product metrics review: summary, scorecard, trend analysis, bright spots, areas of concern, and recommended actions.
What it needs:
- Time period (weekly, monthly, quarterly)
- The metrics and their current values
- Comparison data (previous period, targets)
- Context on recent events (launches, incidents, campaigns)
What it produces:
- Summary (2-3 sentences: overall health and key callout)
- Metric scorecard (table with current/previous/change/target/status)
- Trend analysis per significant metric
- Bright spots and areas of concern
- Recommended actions (specific, not vague)
Worked Example: InsightFlow Monthly Metrics Review
Here is a monthly /metrics-review for InsightFlow using illustrative post-Sprint 1 data:
/metrics-review monthly
Period: February 2026 vs January 2026
Product: InsightFlow (B2B SaaS analytics, Series B)
North Star:
- WAT creating/editing dashboards: 521 teams (Jan: 498) +4.6%
- Target: 550 by end of Q1. Status: At risk (-5% vs target)
L1 Metrics:
Acquisition:
- New trial signups: 340 (Jan: 310) +9.7%
- Signup-to-activation: 34% (Jan: 37%) -3pp ⚠️
Engagement:
- WAU total: 2,104 (Jan: 1,980) +6.3%
- DAU/MAU stickiness ratio: 0.41 (Jan: 0.39)
Retention:
- D30 retention: 61% (Jan: 64%) -3pp ⚠️
- Enterprise account retention: 94% (stable, good)
- Free tier D30 churn: 21% (Jan: 18%) ⚠️
Monetisation:
- MRR: $280K (Jan: $265K) +5.7%
- Free-to-Pro conversion: 8.2% (Jan: 8.8%) -0.6pp
Satisfaction:
- NPS: 38 (Jan: 39) flat
Context:
- Workflow Builder Sprint 1 shipped Feb 14 — trigger config UI
now in design testing, no public availability yet
- Self-serve onboarding redesign launched Jan 31 (partially explains
the Feb activation drop — new flow may have friction)
- No major incidents
Sample output from /metrics-review:
MONTHLY METRICS REVIEW — February 2026
InsightFlow | Period: Feb vs Jan 2026
SUMMARY
Overall product health: mixed. MRR and total active users growing
well. Two concerning signals require attention: activation rate
dropped 3pp (likely related to the onboarding redesign) and free-tier
D30 churn increased 3pp. The North Star is growing but behind target.
METRIC SCORECARD
| Metric | Feb | Jan | Change | Target | Status |
|-------------------------|--------|--------|---------|---------|-------------|
| WAT (North Star) | 521 | 498 | +4.6% | 550 | ⚠️ At risk |
| New signups | 340 | 310 | +9.7% | 320 | ✅ On track |
| Activation rate | 34% | 37% | -3pp | 40% | ❌ Miss |
| WAU | 2,104 | 1,980 | +6.3% | 2,000 | ✅ On track |
| DAU/MAU stickiness | 0.41 | 0.39 | +0.02 | 0.42 | ⚠️ Close |
| D30 retention | 61% | 64% | -3pp | 65% | ❌ Miss |
| Free-tier churn (D30) | 21% | 18% | +3pp | 18% | ❌ Miss |
| Enterprise retention | 94% | 94% | flat | 92%+ | ✅ On track |
| MRR | $280K | $265K | +5.7% | $270K | ✅ On track |
| Free-to-Pro conversion | 8.2% | 8.8% | -0.6pp | 9.0% | ⚠️ At risk |
| NPS | 38 | 39 | -1 | 45 (Q2) | ⚠️ Monitor |
AREAS OF CONCERN
1. Activation rate: 34% vs 37% prior month (-3pp)
Likely cause: The Jan 31 onboarding redesign may have introduced
friction. New users are signing up at a higher rate but activating
at a lower rate — suggesting the top-of-funnel is working but the
activation flow is not.
Action: Pull L2 funnel data — where in the new onboarding flow
are users dropping? Funnel analysis this week.
2. Free-tier D30 churn: 21% vs 18% (+3pp)
Correlates with the activation drop — users who do not activate
churn faster. If activation improves, churn should follow.
Action: Cohort analysis: do Feb signups churn at the same rate
at D7, D14, D30 as Jan signups? Confirm the correlation.
BRIGHT SPOTS
- Enterprise retention at 94% — strong. Consistent with SOC 2 work.
- MRR growth +5.7% — above target.
- Signup volume growing despite activation headwind — top of funnel healthy.
RECOMMENDED ACTIONS
1. [This week] Pull L2 funnel data for new onboarding flow — identify
the specific step where Feb cohort drops below Jan cohort.
2. [This week] Cohort comparison: Jan signups vs Feb signups at D7/D14
— confirm churn signal is real, not seasonal noise.
3. [This sprint] If funnel data confirms onboarding friction — scope
a fix as Sprint 2 P1 item. Do not wait for more churn evidence.
4. [Ongoing] North Star is 4.6% short of Q1 target. Review whether
the Workflow Builder will drive WAT growth when it ships, or
whether a faster initiative is needed.
Evaluating the Output
Before acting on this review, check these quality criteria from the /metrics-review skill:
Does every metric have a comparison? A raw number without context is not information. The scorecard shows current, previous, change, target, and status for every metric. If any metric shows only a current value, prompt: "Add prior period and target to [metric]."
Are anomalies explained? The activation drop and churn increase are flagged with a likely cause (onboarding redesign). The agent attributes them as likely correlated — note the epistemic humility here. Correlation is not causation. The review recommends investigation, not a conclusion.
Do recommended actions name a specific next step? "Investigate further" is not a recommended action. "Pull L2 funnel data for new onboarding flow — identify the specific drop-off step" is.
The metrics review you run in this exercise feeds directly into Lesson 14, where you will run a retrospective on Sprint 1 and evaluate whether the metrics you tracked were the right ones to measure.
Try With AI
Use these prompts in Cowork or your preferred AI assistant.
Prompt 1 — Reproduce (apply what you just learned):
Run a monthly metrics review for InsightFlow using this data:
North Star (WAT creating/editing dashboards): 521 (prior: 498, target: 550)
Activation rate: 34% (prior: 37%, target: 40%)
Free-to-Pro conversion: 8.2% (prior: 8.8%, target: 9.0%)
D30 retention: 61% (prior: 64%, target: 65%)
MRR: $280K (prior: $265K, target: $270K)
NPS: 38 (prior: 39, target Q2: 45)
Context: New onboarding flow launched Jan 31 — may explain activation
drop. No major incidents. Workflow Builder Sprint 1 shipped Feb 14
but no public availability yet.
Produce: summary (3 sentences), metric scorecard, 2 areas of concern
with likely causes, 3 recommended actions.
What you're learning: Running the full /metrics-review workflow with structured data. The key skill is evaluating whether the agent's "likely causes" are well-reasoned or speculative — and prompting for more rigour when needed.
Prompt 2 — Adapt (change the context):
Write OKRs for InsightFlow Q2 based on the February metrics review.
The February review identified these gaps:
- Activation rate 34% vs 40% target (onboarding friction)
- Free-tier D30 churn 21% vs 18% target
- North Star growing slowly vs Q1 target
Write one Objective and 3-4 Key Results for Q2 that address these gaps.
Rules:
- Each KR must be measurable with a specific metric, baseline, and target
- No KR may measure team output (features shipped, tickets completed)
- 70% completion should feel ambitious but achievable
Then evaluate your own KRs: are any of them output-based rather
than outcome-based? If so, rewrite them.
What you're learning: The OKR self-evaluation exercise — writing KRs and then testing them against the output-vs-outcome rule. This is the skill that separates OKRs that drive behaviour from OKRs that just look good in a planning doc.
Prompt 3 — Apply (connect to your domain):
Define your product's metrics hierarchy:
1. North Star: What single action, when a user takes it, means they
are getting real value from your product? Write it as a measurable
metric (e.g., "weekly active teams creating or editing dashboards").
Explain why you chose this metric over alternatives like MAU or revenue.
2. L1 Health Indicators: List 5-6 metrics that together cover:
Acquisition | Activation | Engagement | Retention | Monetisation | Satisfaction
For each, state: the metric name, current value (or "unknown"), and target.
3. Dashboard anti-patterns: Look at your current product dashboard.
Are any metrics vanity metrics (always go up regardless of product quality)?
Is any metric showing a raw number without a comparison? Name and fix it.
What you're learning: Building your own metrics hierarchy. The North Star definition forces clarity on what "value delivered" means in your product — a question many PMs have never explicitly answered.
Exercise: InsightFlow Metrics Review + OKRs
Plugin: Official product-management
Command: /metrics-review
Time: 30 minutes
Step 1 — Define InsightFlow's North Star and L1 metrics
Before running /metrics-review, write down:
- InsightFlow's North Star metric (from the spec above: WAT creating/editing dashboards)
- Why this is better than MAU or revenue
- The 5 L1 metrics you will track in the review
This step ensures you are reviewing the right things, not just whatever data is available.
Step 2 — Run /metrics-review with sample data
/metrics-review monthly
Period: February 2026 (4 weeks post-sprint 1 start)
North Star — WAT (weekly active teams creating/editing dashboards):
Current: 521 | Prior month: 498 | Target: 550
Activation — % signups creating first dashboard within 7 days:
Current: 34% | Prior: 37% | Target: 40%
Retention — D30 retention rate:
Current: 61% | Prior: 64% | Target: 65%
Monetisation — Free-to-Pro conversion:
Current: 8.2% | Prior: 8.8% | Target: 9.0%
Acquisition — New trial signups:
Current: 340 | Prior: 310 | Target: 320
Satisfaction — NPS:
Current: 38 | Prior: 39 | Target Q2: 45
Context:
- New onboarding flow launched Jan 31 — first month of data
- Workflow Builder Sprint 1 shipped Feb 14, design testing only
- No major incidents or marketing campaigns
Step 3 — Evaluate the output
Check these criteria:
- Does every metric in the scorecard show current, prior, change, target, and status? If not, prompt for the missing comparisons.
- For the activation drop: does the agent propose a likely cause? Is that cause reasonable given the context (new onboarding flow)? Or is it speculative?
- Are the recommended actions specific? "Investigate activation further" is too vague. "Pull funnel drop-off data for the new onboarding flow by step" is specific.
Step 4 — Write OKRs for next quarter
Based on the metrics review, write one Objective and 3-4 Key Results for InsightFlow Q2:
Draft InsightFlow Q2 OKRs based on this metrics review:
The gaps are:
- Activation at 34% vs 40% target
- Free-tier churn at 21% vs 18% target
- North Star growing slower than Q1 target
Write one Objective (aspirational, qualitative) and 3-4 Key Results
(measurable, outcome-based, ambitious but achievable at 70%
confidence). Each KR must name a specific metric, current baseline,
and Q2 target.
Step 5 — Evaluate your OKRs for the output trap
For each Key Result, ask: "Could the team hit this KR without a single user benefiting?" If yes — if the KR could be hit by shipping a feature that no one uses — rewrite it to measure user behaviour instead.
Save the OKR document. It becomes the strategic backdrop for the retrospective in Lesson 14.
What You Built
You built InsightFlow's product metrics hierarchy — North Star, five L1 health indicators, and an understanding of when to use L2 diagnostic metrics. You ran a monthly metrics review against illustrative post-Sprint 1 data and produced a scorecard with trend analysis, areas of concern, and three specific recommended actions.
You also wrote Q2 OKRs using the output-versus-outcome discipline, ensuring your Key Results measure user behaviour rather than team activity.
The metrics review feeds directly into Lesson 14, where you will retrospect on Sprint 1 and evaluate whether the success metrics you tracked were the right ones — and whether the team built what you intended.
Flashcards Study Aid
Continue to Lesson 14: Continuous Intelligence — Agents & Retrospectives →