Metric Glossary
A definitive handbook of product metrics. Understand the math, the meaning, and the common pitfalls before you optimize.
106 metrics covering Growth, Revenue, Retention, and Strategy.
DAU
growthDaily Active Users
Count(Unique Users with > 0 Sessions in 24h)- Avoid vanity: Define "Active" strictly. Viewing a page might not be enough; performing a core action is better.
- Fluctuations: DAU can be volatile on weekends vs weekdays.
Why this metric?
MAU
growthMonthly Active Users
Count(Unique Users with > 0 Sessions in 30 days)- Lagging indicator: A drop in MAU is often noticed too late.
- Good for high-level health checks, but less actionable than DAU or WAU.
WAU
growthWeekly Active Users
Count(Unique Users with > 0 Sessions in 7 days)- Ideal for B2B products where usage is consistently weekly but not necessarily daily.
- Less volatile than DAU, more actionable than MAU.
Active Hosts
growthActive Hosts (Marketplace)
Count(Hosts with > 0 Active Listings or Sessions)- Supply constraints: In marketplaces, supply is often the constraint. Tracking active hosts is as critical as active buyers.
- Quality over Quantity: 10 active super-hosts are better than 100 inactive ones.
New Host Signups / Signups
growthNew User Signups
Count(Users Registering Account)- Top of Funnel: High signups with low activation is a sign of poor targeting or onboarding, not success.
- Spam Risk: Be aware of bot signups inflating this number.
Activation Rate
growthUser Activation Rate
(Users who performed Activation Event / Total New Users) * 100- Critical metric: Signups without activation is a leaky bucket.
- Time-bound: Always define a window (e.g. Activated within 7 days).
Why this metric?
Subscriber Base
growthTotal Subscribers
Count(Users with status = "active")- Paused vs Churned: Decide how to handle paused subscriptions. They are not active, but not fully churned.
- Tier Mix: Track the breakdown of Basic vs Premium subscribers.
Organic Traffic
growthOrganic User Traffic
Count(Sessions where Source = Organic Search)- SEO Health: Primary indicator of SEO performance.
- Brand Strength: High direct/organic traffic often signals strong brand recall.
Viral Coefficient (K-factor)
growthViral Coefficient
Avg # of Invitations sent * Conversion rate of Invitations- K > 1: Exponential growth (Viral).
- K < 1: Linear or decaying growth (Paid/Organic dependent).
CPI
growthCost Per Install
Total Ad Spend / Total Installs- Platform Variance: CPI on iOS is typically higher than Android.
- Fraud: Watch out for install farms inflating numbers.
PMF
generalProduct-Market Fit
Qualitative (Sean Ellis Test: >40% would be "very disappointed" without it)- Sustainable growth: PMF is the precursor to scaling. Don't pour fuel on a product without PMF.
- Retention as proxy: High long-term retention is the best quantitative indicator of PMF.
Why this metric?
Aha! Moment
growthAha! Moment
User performs [Action] + [Frequency] within [Time Period]- Facebook: 7 friends in 10 days.
- Slack: 2,000 messages sent within a team.
- Dropbox: 1 file in 1 folder on 1 device.
Referral Rate
growthReferral Rate
(Number of Referrals / Total Customers) * 100- Advocacy: High referral rate indicates strong product love.
- Incentives: Track organic vs incentivized referrals separately.
PQL
growthProduct Qualified Leads
Users hitting Activation criteria- Sales alignment: PQLs convert much higher than MQLs (Marketing Qualified Leads).
- Usage based: Defined by actual product usage, not just demographic fit.
ARR
revenueAnnual Recurring Revenue
Sum(Annual Subscription Value of all active customers)- Don't include one-time fees (consulting, setup).
- Committed vs Contracted: Use committed revenue for accurate forecasting.
GMV
revenueGross Merchandise Value
Sales Price * Number of Items Sold- Revenue is separate: GMV is not your revenue; it is the throughput. Revenue is GMV * Take Rate.
- Returns: Gross GMV usually includes returns; Net GMV excludes them.
GMS
revenueGross Merchandise Sales
Sum(Sales Volume)- Marketplace Context: Often used interchangeably with GMV but specifically refers to the sales volume processed.
- Net GMS: Sales after cancellations and returns.
Take Rate
revenueTake Rate / Commision Rate
(Revenue / GMV) * 100- Marketplace Health: High take rate could drive sellers away; low take rate might make the business unsustainable.
- Components: Often includes commission + payment fees + advertising revenue.
TPV
revenueTotal Payment Volume
Sum(Value of all processed transactions)- Metric for Fintech: Similar to GMV for e-commerce. Indicates scale/throughput.
- Currency fluctuations: Be careful when aggregating TPV across multiple currencies.
AOV
revenueAverage Order Value
Total Revenue / Total Number of Orders- Pricing Strategy: Increasing AOV is often easier than getting new customers (bundling, upsells).
- Beware of outlines: One whale buyer can skew the average. Check the median too.
ARPU
revenueAverage Revenue Per User
Total Revenue / Total Active Users- Segementation key: ARPU varies wildly by cohort (e.g. Free vs Pro). Always segment.
- Vanity warning: High ARPU with shrinking user base is a warning sign of a niche product.
LTV
revenueLifetime Value
(ARPU * Gross Margin %) / Churn Rate- Prediction hazard: LTV is a forecast, not a fact. Be conservative with churn assumptions.
- LTV:CAC Ratio: Ideally 3:1. If 1:1, you are bleeding money. If 5:1, you are under-investing in growth.
Why this metric?
CAC
revenueCustomer Acquisition Cost
(Sales + Marketing Expenses) / Number of New Customers Acquired- Blended vs Paid: Blended CAC hides the inefficiency of paid channels. Always calculate Paid CAC separately.
- Time lag: Marketing spend today might not produce users for months (in B2B).
Why this metric?
LTV:CAC
revenueLTV to CAC Ratio
LTV / CAC- Benchmark: 3:1 is the gold standard for SaaS.
- Efficiency: If > 5:1, you are likely under-spending on growth.
CLV
revenueCustomer Lifetime Value
Avg Order Value * Purchase Frequency * Customer Lifespan- CLV vs LTV: LTV often refers to Gross Margin contributed, while CLV can refer to pure revenue.
- Cohort tracking: Track CLV by signup month to see if customer quality is improving.
CAC Payback Period
revenueCAC Payback Period
CAC / (ARPU * Gross Margin %)- Cash efficiency: For startups, < 12 months is the goal. > 18 months creates cash flow drag.
- Segment: SMB payback should be faster (6-9mo) than Enterprise (12-18mo).
Gross Margin
revenueGross Margin
((Revenue - COGS) / Revenue) * 100- SaaS scalability: High gross margins (80%+) explain why software companies are valued highly.
- COGS: Includes hosting, support, and payment fees, but not R&D.
EBITDA
revenueEBITDA
Net Income + Interest + Taxes + Depreciation + Amortization- Valuation: Common valuation metric for mature companies.
- Rule of 40: Growth Rate + EBITDA Margin should be > 40%.
Win Rate
revenueSales Opportunity Win Rate
(Closed Won Deals / Total Opportunities) * 100- Pipeline Quality: Low win rate usually means poor lead qualification, not just bad sales skills.
- Stage analysis: Track conversion rate per stage to find the bottleneck.
Pipeline Value
revenueSales Pipeline Value
Sum(Opportunity Value * Probability to Close)- Forecast accuracy: Weighted pipeline is only as good as the probability estimates.
- Stale deals: Remove deals that have been "open" too long to keep pipeline realistic.
New Business Revenue
revenueRevenue from New Customers
Sum(Revenue from Customers with StartDate in Period)- Hunter vs Farmer: Measures the effectiveness of your sales "hunters".
- Discounting impacts: Watch out for high new biz rev driven by unsustainable discounts.
Expansion Revenue
revenueExpansion Revenue
Sum(Upsell + Cross-sell Revenue)- Cheaper growth: Expanding existing customers is 5-25x cheaper than acquiring new ones.
- Net Retention driver: This is the fuel for NDR > 100%.
Burn Rate
revenueNet Burn Rate
Cash Spent - Cash Revenue (per month)- Runway calculator: Cash Balance / Burn Rate = Months of Runway.
- Efficiency: High burn is acceptable only if growth is efficiently high (Magic Number > 1).
Runway
revenueCash Runway
Current Cash Balance / Monthly Net Burn Rate- Survival metric: 18-24 months is standard for venture-backed startups.
- Fundraising trigger: Usually need to start raising when runway < 6-9 months.
Magic Number
revenueSaaS Magic Number
(Current Q Revenue - Previous Q Revenue) * 4 / Previous Q Sales & Marketing Spend- > 1.0: Efficient growth. Pour more fuel on the fire.
- < 0.7: Inefficient. Fix the funnel before spending more.
MRR
revenueMonthly Recurring Revenue
Sum(Monthly Subscription Fees)- Momentum: The most important metric for SaaS growth.
- Commitment: Includes recurring charges, excludes one-time fees.
Why this metric?
Net Revenue Churn
revenueNet Revenue Churn
((Revenue Lost from Churn - Expansion Revenue) / Starting Revenue) * 100- Negative Churn: If Expansion > Churn, you have Negative Net Churn (the holy grail).
- Sustainability: High net churn kills growth.
Quick Ratio
revenueSaaS Quick Ratio
(New MRR + Expansion MRR) / (lost MRR + Contraction MRR)- > 4: Excellent growth efficiency.
- < 2: You are burning cash just to replace lost customers.
Rule of 40
revenueRule of 40
Annual Revenue Growth Rate % + EBITDA Margin %- Trade-off: You can be unprofitable (-10%) if you are growing fast (50%).
- Balance: High growth permits lower margins, and vice versa.
Burn Multiple
revenueBurn Multiple
Net Burn / Net New ARR- < 1.0: Amazing efficiency.
- > 3.0: Warning sign, spending too much for too little growth.
Churn Rate
retentionCustomer Churn Rate
(Lost Customers / Total Customers at Start of Period) * 100- Silent killer: Low churn is often more important than high growth for sustainability.
- Revenue Churn vs Logo Churn: You can lose customers but grow revenue (negative churn) if upsells are strong.
Why this metric?
NDR / NRR
retentionNet Dollar Retention
((Review Start Rev + Expansion - Churn - Contraction) / Start Rev) * 100- >100% is the holy grail: It means you grow even if you acquire zero new customers.
- Cohort analysis: Analyze NDR by cohort to see if older cohorts expand or shrink over time.
Why this metric?
Repeat Purchase Rate
retentionRepeat Customer Rate
(Customers with > 1 Order / Total Unique Customers) * 100- E-commerce vital sign: Breaking even on first purchase is rare; profit comes from repeats.
- Window matters: Define the window carefully (e.g. repeat within 90 days).
Stickiness
retentionStickiness Ratio (DAU/MAU)
(DAU / MAU) * 100- World class: > 20% is good; > 50% is Facebook level.
- Utility drift: Only relevant for daily-use products.
Sessions per User
engagementAverage Frequency of Use
Total Sessions / Total Active Users- Habit Strength: High frequency often indicates strong habit formation.
- Natural Frequency: Don't expect daily use for a monthly utility app (e.g. Payroll).
Time Spent
engagementAverage Time Spent per User
Total Session Duration / Total Active Users- Quality vs Quantity: Time spent is good for media (Netflix), but bad for utility (Uber).
- Active vs Idle: Ensure you are tracking active foreground time.
Content Consumption
engagementContent Consumption
Sum(Content Units Consumed)- Feed Health: Critical for media/social apps.
- Completion Rate: Watching 5% of a video counts as a view, but completion is a better signal of quality.
Content Engagement
engagementEngagement with Content
Sum(Likes + Comments + Shares)- Algorithm signal: These are high-value signals for recommendation algorithms.
- Share > Like: A share is usually a much stronger endorsement than a like.
Content Creation / Files Created
engagementUser Generated Content Volume
Count(New Core Entities Created)- Prosumer Activity: Creation is often the first step to collaboration/sharing.
- Empty states: Watch out for "Created but empty" files.
Messages Sent
engagementCommunication Volume
Count(Messages Successfully Sent)- Network Density: More messages usually means a denser, stickier network.
- Bot traffic: Exclude automated system messages.
Project Activity
engagementProject Management Activity
Count(Task Updates + Comments + Transitions)- Work vs Noise: Are users just moving cards around, or actually closing them?
- Collaboration proxy: High activity implies the team is living in the tool.
Meeting Volume
engagementMeeting Activity
Count(Meetings with > 1 participant)- Cost metric: For internal tools, high meeting volume might be bad (inefficiency).
- Value metric: For Zoom/Teams, this is the core value unit.
Bounce Rate
engagementWebsite Bounce Rate
(Single Page Sessions / Total Sessions) * 100- Content relevance: High bounce rate often means the landing page didn't match the ad promise.
- Exceptions: Single-page apps (SPAs) or simple informational queries might have naturally high bounce rates.
Cart Abandonment
engagementCart Abandonment Rate
((Initiated Carts - Completed Transactions) / Initiated Carts) * 100- Friction point: Often caused by surprise costs (shipping, tax) at checkout.
- Retargeting: Primary trigger for abandoned cart emails.
NPS
generalNet Promoter Score
% Promoters (9-10) - % Detractors (0-6)- Sentiment vs Reality: Users may say they love you (high NPS) but still churn.
- Use strictly as a pulse check, not a root cause diagnostic tool.
Conversion Rate
generalFunnel Conversion Rate
(Users entering Stage B / Users entering Stage A) * 100- Micro-conversions: Don't just track the final sale; track each step to find the friction.
- Local maxima: optimizing one step (e.g. clicks) might hut the next down stream retention.
Resolution Time
generalAverage Resolution Time
Average(Resolved Date - Created Date)- Customer Satisfaction (CSAT): Long resolution times kill CSAT.
- SLA Breaches: Track % of tickets resolved within SLA, not just average.
Cycle Time
generalDevelopment Cycle Time
Average(Completion Timestamp - Start Timestamp)- Flow efficiency: High cycle time often indicates blocking dependencies or context switching.
- Predictability: Stable cycle time is better than fast but erratic cycle time.
User Efficiency
generalUser Workflow Efficiency
Average(Time to Complete Core Action)- Friction hunting: Use this to identify UX hurdles.
- Learning curve: Efficiency should improve as user tenure increases.
Supply Liquidity
generalMarketplace Supply Liquidity
(Transactions / Active Listings) * 100- Marketplace Health: High liquidity means buyers always find what they want.
- Cold start: Hardest metric to move in a new marketplace.
SUS
generalSystem Usability Scale
Survey Score (0-100)- Industry standard: Score > 68 is considered above average.
- Quick pulse: "I thought the system was easy to use."
CES
generalCustomer Effort Score
Survey: "How easy was it to handle your issue?"- Loyalty predictor: High effort correlates strongly with churn.
- Frictionless: The goal is to make interactions effortless.
Task Success Rate
generalTask Success Rate
(Successful Completions / Total Attempts) * 100- Usability validation: If users can't finish the task, the design failed.
- Severity: Distinguish between "gave up" vs "completed with errors".
Time on Task
generalTime on Task
End Time - Start Time- Efficiency: Generally lower is better for utility tasks.
- Benchmarking: Compare against expert users or previous versions.
NSM
generalNorth Star Metric
Context Dependent (Unique to each product)- Not Revenue: Revenue is a lagging indicator of value. The North Star should reflect value received.
- Focus alignment: It serves to align the entire company towards a singular product goal.
Input Metric
generalInput Metric
Actionable & Controllable by teams (e.g. Inventory growth, Latency reduction)- Actionable: Teams should be able to influence these directly.
- Hypothesis-driven: You believe moving these will eventually move the North Star.
Output Metric
generalOutput Metric
Result-oriented (e.g. Revenue, Total Churn)- Lagging: By the time you see them move, the activities that caused the move are already over.
- Hard to influence directly: You move them by moving input metrics.
Counter Metric
generalCounter Metric
Dependency check (e.g. If NSM is Orders, Counter might be Order Cancellations)- Checks & Balances: Prevents teams from "gaming" the system (e.g. increasing signups but hurting quality).
- Health check: If the counter metric tanks, your NSM growth might be hollow.
OKR
generalObjectives and Key Results
Objective (Vision) + 3-5 Key Results (Measurable Metrics)- Ambitious: Objectives should be qualitative and inspiring.
- Quantitative: Key Results must be numbers (e.g. Grow MAU from 1M to 1.5M).
KPI
generalKey Performance Indicator
Business Vital Signs- Steady state: KPIs are often monitored continuously to ensure business health.
- NSM vs KPI: A North Star is a type of KPI, but usually the most important one.
MVP
generalMinimum Viable Product
Leanest version for Learning- Learning tool: The goal is to test hypotheses, not just build a "cheap" version.
- Viable: It must still solve the core problem for the user.
PRD
generalProduct Requirements Document
The "What" and "Why" of a feature/product- Alignment: Ensures that everyone on the team knows what is being built and for whom.
- Living document: Should be updated as requirements evolve.
User Persona
generalUser Persona
Demographics + Behaviors + Pain Points + Goals- Empathy: Helps the team design with a specific human in mind.
- Segmentation: Different personas might have different North Star Metrics if they find different value.
Adoption Rate
growthFeature Adoption Rate
(Users of Feature / Total Eligible Users) * 100- Discovery vs Utility: Low adoption could mean they can't find it (Discovery) or they don't need it (Utility).
- Depth: Adoption is binary; also track frequency/depth of use.
TTV
growthTime to Value
Average(Activation Timestamp - Signup Timestamp)- Onboarding velocity: Faster TTV usually leads to higher retention.
- Complex products: For B2B, TTV might be measured in weeks; for B2C, in minutes.
Guardrail Metric
generalGuardrail Metric
Latency, Error Rate, Cancellation Rate- Safety first: If a guardrail metric violates a threshold, stop the experiment immediately.
- Counter Metric vs Guardrail: Counter metrics are often business-oriented (Profit vs Revenue); Guardrails are often technical (Latency, Crashes).
Lead Time for Changes
generalLead Time for Changes
Deployment Timestamp - Commit Timestamp- Agility: Lower lead time means faster feedback loops.
- Bottlenecks: Highlights delays in QA or CI/CD pipelines.
MTTR
generalMean Time to Recovery
Sum(Down time) / Count(Incidents)- Resilience: It is not about never failing, but recovering fast.
- SLA: Critical for enterprise contracts.
Match Rate
generalMatch Rate
(Successful Matches / Total Searches or Requests) * 100- Liquidity quality: A high match rate ensures users don't leave empty-handed.
- Zero results: Track "Zero Result Searches" closely.
Search to Fill
generalSearch to Fill Rate
(Transactions / Total Searches) * 100- Relevance: Indicates how well the inventory matches user intent.
- Pricing: Also affected by price matching.
Buyer/Seller Overlap
generalBuyer/Seller Overlap
(Users who Buy AND Sell / Total Users) * 100- Network effects: High overlap (e.g. Poshmark, Airbnb) creates a powerful, self-sustaining ecosystem.
- Acquisition: Acquiring one user gets you both supply and demand.
Crash-Free Users
generalCrash-Free Users Rate
((Total Users - Users with Crash) / Total Users) * 100- Stability: Target > 99.9% for high-quality apps.
- Retention killer: Crashes are the fastest way to lose a mobile user.
ANR Rate
generalApplication Not Responding Rate
(ANR Sessions / Total Sessions) * 100- Frustration: Worse than a crash because the user waits.
- Performance: Often tied to main thread blocking.
App Store Conversion
generalApp Store Conversion Rate
(Installs / Page Views) * 100- ASO: Optimize screenshots, reviews, and description.
- First impression: This is the landing page for mobile.
DAP / MAP
growthDaily/Monthly Active People
Count(Unique Physical Humans Active)- De-duplication: Requires advanced identity matching to link devices/accounts.
- Family Metrics: Used to aggregate usage across a family of apps (FB, Insta, WhatsApp).
Resurrection Rate
retentionUser Resurrection Rate
(Resurrected Users / Total Dormant Users) * 100- Win-back: Measures the effectiveness of re-engagement campaigns (email, push).
- Cheaper than new: Resurrecting a user is often cheaper than acquiring a completely new one.
Fulfillment Rate
generalFulfillment Rate
(Completed Orders / Total Orders Placed) * 100- Marketplace Trust: Low fulfillment kills trust on the demand side.
- Supply issue: Often indicates lack of supply liquidity or operational failures.
Streak
engagementUser Streak Length
Count(Consecutive Active Periods)- Gamification: The most powerful retention mechanic in EdTech (Duolingo) and Health.
- Loss Aversion: Users return just to "save the streak".
KYC Success Rate
generalKYC Pass Rate
(Verified Users / Total KYC Attempts) * 100- Funnel Blocker: KYC is the biggest drop-off point in Fintech onboarding.
- Fraud balance: Too easy = fraud; Too hard = lost users.
Items per Basket
revenueItems Per Basket (UPT)
Total Items Sold / Total Transactions- Bundling proxy: High UPT usually means successful cross-selling.
- Inventory efficiency: Helps move more SKUs per logistics cost unit.
Unit Sales
revenueUnit Sales
Count(Items Sold)- Hardware North Star: For Apple/Tesla, units sold is the primary measure of scale.
- Install Base: Units sold accumulates into Active Installed Base.
Lessons Completed
engagementLessons Completed
Count(Completed Lessons)- Value metric: For EdTech, this is the core value exchange.
- Paywall trigger: Often used as a limit before requiring payment.
Online Hours
growthProvider Online Hours
Sum(Time Available Online)- Supply Capacity: The raw capacity of a service marketplace.
- Peak vs Off-peak: Utilization rate is Online Hours / Busy Hours.
Health Logs
engagementActivities/Meals Logged
Count(User Logs)- Active Input: Requires high effort; indicators of high intent.
- Retention: Logging is usually the primary habit loop in health apps.
Velocity
generalSprint Velocity
Sum(Story Points of Completed User Stories in a Sprint)- Not a comparison tool: Velocity is unique to each team. Do not compare Team A's velocity to Team B's.
- Predictability: The goal is a stable velocity, not an constantly increasing one.
Lead Time
generalProduct Lead Time
Average(Delivery Timestamp - Creation Timestamp)- Customer Centric: This is the time the customer actually experiences waiting.
- Cycle vs Lead: Cycle time is when dev starts; Lead time is from when the ticket is created.
WSJF
generalWeighted Shortest Job First
Cost of Delay / Job Size- Cost of Delay = User-Business Value + Time Criticality + Risk Reduction/Opportunity Enablement.
- Job Size = Effort/Duration. Doing high-value, easy things first wins.
TTV
engagementTime To Value
Average(Time of Aha Moment - Time of Signup)- Onboarding: Shortening TTV is the primary goal of user onboarding.
- B2B vs B2C: TTV in B2C is often measured in minutes; in Enterprise B2B, it can be weeks/months.
GRR
retentionGross Retention Rate
((Starting Revenue - Revenue Lost to Churn & Downsells) / Starting Revenue) * 100- Unlike NDR, GRR cannot exceed 100%. It tells you the absolute health of your base retention.
- Benchmark: > 90% is excellent for SaaS.
Free-to-Paid Conversion
revenueFree-to-Paid Conversion Rate
(Number of Converted Paid Users / Total Free Users in Cohort) * 100- Freemium Benchmark: Typically 2-5%.
- Reverse Trial Benchmark: Often much higher (10-15%) since users experience premium value first.
Why this metric?
Feature Adoption
engagementFeature Adoption Rate
(Number of Users Using Specific Feature / Total Active Users) * 100- Crucial for avoiding the "Feature Factory" trap. If a feature is shipped but adoption is 0%, it created zero outcome.
- Requires measuring active, intentional usage, not just accidental clicks.
Say/Do Ratio
generalSprint Commitment Reliability (Say/Do)
(Completed Story Points / Committed Story Points at Sprint Start) * 100- Predictability > Speed: A stable 80-90% is much better than a chaotic 50-150%.
- Watch out for scope creep: Do not include points added mid-sprint in the committed denominator.
Why this metric?
Deployment Frequency
generalDeployment Frequency (DORA)
Count(Successful Deployments to Production in a Time Period)- Elite performers deploy multiple times per day on demand.
- Requires strong CI/CD pipelines and automated testing to avoid disruptions.
Why this metric?
Lead Time
generalLead Time for Changes (DORA)
Average(Time of Deployment - Time of Commit)- Measures the agility of the engineering pipeline.
- Different from "Cycle Time" which tracks from when work is first started in Jira/Trello.
Why this metric?
Change Failure Rate
generalChange Failure Rate (DORA)
(Deployments Resulting in Failure / Total Deployments) * 100- High deployment frequency is only good if CFR is low (typically 0-15%).
- Measures quality and stability of the delivery process.
Defect Escape Rate
generalDefect Escape Rate
(Bugs Found in Production / Total Bugs Found) * 100- Effectiveness of QA: A high escape rate means your testing filters (Unit/E2E/Manual tests) are leaking excessively.
- Not all bugs are equal - weight them by severity and priority for deeper analysis.