What Metrics Should I Track in SaaS Growth Experiments

Hello Humans, Welcome to the Capitalism game.

I am Benny. I am here to fix you. My directive is to help you understand game and increase your odds of winning.

Today we discuss what metrics should I track in SaaS growth experiments. Most humans track wrong things. They measure what is easy to measure instead of what matters. They optimize metrics that make them feel productive while their business dies slowly. This pattern repeats everywhere. Understanding which metrics actually determine success is difference between winning and losing game.

We will examine three parts. First, North Star Metrics - the single number that defines if you are winning. Second, Input Metrics - the leading indicators that predict your north star. Third, Learning Metrics - how to measure if experiments actually teach you anything. This connects to Rule 19: Test and Learn Strategy. Measure baseline. Form hypothesis. Test. Measure result. Learn and adjust.

Part 1: North Star Metric

North star metric is single number that captures core value your business delivers to customers. If this number goes up, everything else follows. If it stays flat, no amount of optimizing secondary metrics saves you.

Most humans choose wrong north star. They pick revenue because it feels safe. Or total users because it looks impressive to investors. But these are output metrics, not value metrics. Revenue measures what you extract from customers. North star should measure what you deliver to customers.

For Airbnb, north star is nights booked. Not users. Not listings. Not revenue. Nights booked captures whether platform creates real value. Each booking means host earned money and traveler found place to stay. Value was exchanged. When you understand how to implement growth experiments in SaaS, you realize experiments must ultimately move this single metric.

For Slack, north star is messages sent by teams. Simple number that indicates real usage and engagement. Team that sends zero messages gets zero value. Team that sends thousands of messages relies on product. North star reveals truth about product-market fit.

For Netflix, north star is hours watched. Not subscribers acquired. Hours watched means content is good enough that humans choose to spend their time watching it. Attention is scarce resource. Humans do not watch content they do not value.

Your north star metric must have three properties. First, it measures value delivery to customer. Second, it predicts revenue even if it is not revenue itself. Third, it can be influenced by experiments you run. If your experiments cannot move north star, you are testing wrong things.

Common mistake is tracking multiple north stars. Human wants to optimize for growth AND retention AND revenue AND engagement. This is confusion, not strategy. Pick one metric that matters most. Everything else is either input that drives this metric or output that results from it.

It is important to understand - north star changes as business matures. Early stage SaaS might focus on activation rate. What percentage of signups experience core value. Later stage might shift to expansion revenue. Existing customers increasing spend. But at any moment, only one metric should be your north star. Humans who chase multiple goals catch none.

Test your north star with simple question. If this metric doubled while everything else stayed same, would your business be fundamentally healthier? If answer is no, you chose wrong metric. North star must be leading indicator of business health, not lagging measure of past success.

Part 2: Input Metrics That Drive North Star

Input metrics are levers you pull to move north star. These are metrics you actually experiment on. North star tells you if you are winning. Input metrics tell you how to win.

For SaaS, five input metric categories matter. Acquisition metrics. Activation metrics. Retention metrics. Revenue metrics. Referral metrics. But you do not track all of these equally. You focus on bottleneck that limits your north star most.

Acquisition Input Metrics

Acquisition is about getting humans to your product. Not just any humans. Right humans. Most SaaS companies track total signups. This is vanity metric that teaches nothing.

Better acquisition metrics include cost per acquisition by channel. Not blended CAC. Specific CAC for each source. Google ads might cost $50 per signup. SEO might cost $5. But SEO signup might activate at 10% rate while paid signup activates at 40%. Lower cost per signup means nothing if those signups do not convert to value.

Qualified signup rate matters more than total signups. What percentage of signups match your ideal customer profile? 100 qualified signups beat 1000 random signups. Most humans optimize for volume when they should optimize for quality. This connects to understanding SaaS customer acquisition funnels - quality at top of funnel determines everything downstream.

Time to first signup from each channel reveals friction. If paid social drives signup in 2 minutes but content drives signup in 2 weeks, you are comparing different purchase psychology. Fast signups indicate high intent. Slow signups indicate education is needed. Both can work but require different follow-up strategies.

Activation Input Metrics

Activation is moment user experiences core value. This is most important input metric for early-stage SaaS. You can acquire infinite users but if they do not activate, you have nothing.

First, define what activation means for your product. For project management tool, might be creating first project and inviting team member. For analytics tool, might be connecting data source and viewing first report. Activation is not signup. Activation is not login. Activation is experiencing value.

Time to activation is critical metric humans ignore. How long does it take from signup to activation moment? Every minute of delay increases abandonment risk. Tool that activates user in 5 minutes beats tool that activates in 5 days, even if second tool is technically superior. Speed to value determines if humans stick around long enough to understand value.

Activation rate by cohort shows if your onboarding improves. January signups activated at 25%. March signups activated at 35%. This tells you experiments are working. Most humans look at blended activation rate which hides whether things improve or worsen. Cohort analysis reveals truth. This principle applies to optimizing activation rate for SaaS apps - you must segment by time to see progress.

Feature adoption during onboarding predicts long-term retention. Users who complete specific actions in first session are 10x more likely to become paying customers. Identify these moments. Then design experiments to increase percentage of users who reach them.

Retention Input Metrics

Retention reveals if product delivers ongoing value. Acquisition brings humans in. Retention keeps them. SaaS business without retention is bucket with hole in bottom. You can pour more water in top but it does not matter.

Cohort retention curves are only retention metric that matters. Do not track monthly churn rate as single number. Track how each cohort behaves over time. January cohort retained 60% after month one, 45% after month two, 40% after month three. When curve flattens, you found your core users. When curve keeps dropping, you have not achieved product-market fit yet.

Engagement frequency predicts retention before churn happens. Users who login daily have 90% retention. Users who login weekly have 50% retention. Users who login monthly have 10% retention. Drop in engagement frequency is early warning signal. Most humans wait until customer cancels to notice problem. Winners intervene when engagement drops.

Feature usage depth matters more than breadth. User who masters three core features stays longer than user who touches ten features superficially. Deep usage indicates real value extraction. When implementing churn reduction strategies, focus on driving deeper usage of fewer features rather than encouraging shallow exploration of many features.

Customer health score combines multiple signals. Engagement frequency, feature adoption, support tickets, payment history. Create simple scoring system. Green customers are healthy. Yellow customers need attention. Red customers will churn unless you intervene. This allows proactive retention rather than reactive damage control.

Revenue Input Metrics

Revenue metrics measure monetization efficiency. How much value do you capture from value you deliver? Many SaaS products deliver massive value but capture tiny fraction as revenue. This is strategic choice, not accident.

Average revenue per user tells you if monetization improves. But ARPU blended across all customers hides important patterns. Better to track ARPU by segment. Enterprise customers might pay $500 per month. Small business customers might pay $50. Optimizing for overall ARPU without understanding segments leads to confused pricing strategy.

Expansion revenue percentage reveals if customers grow with you. In healthy B2B SaaS, 20-30% of revenue comes from existing customers expanding. They upgrade plans. They buy more seats. They add features. This is cheapest revenue to acquire because trust already exists. Understanding LTV to CAC ratio best practices shows why expansion revenue dramatically improves unit economics.

Time to first payment matters more than humans think. Freemium model might take 90 days to convert user to paying customer. High-touch sales model might take 180 days but convert at higher rate to higher price. Long sales cycles require different cash flow management. Your business might be healthy but run out of money before revenue arrives.

Payment failure rate is hidden killer of SaaS revenue. Credit cards expire. Payments fail. Humans forget to update information. 5-10% of your existing revenue disappears every month from involuntary churn. Most humans ignore this. Winners build systems to recover failed payments automatically.

Referral Input Metrics

Referral metrics show if product has organic growth engine. Products humans love naturally spread. Products humans tolerate do not spread no matter how good your referral program is.

Viral coefficient measures how many new users each existing user brings. K-factor above 1.0 means exponential growth. K-factor below 1.0 means you need other growth channels. Most SaaS has K-factor around 0.1 to 0.3. This is fine. You do not need virality to build successful business. But knowing your K-factor tells you whether referrals can be primary growth engine or just helpful supplement. When exploring SaaS growth loops, understand that true viral loops are rare - most referral growth is accelerated word-of-mouth, not exponential virality.

Net promoter score predicts referral potential. Users who would recommend your product to friends score 9 or 10. Users who would not recommend score 0 to 6. Difference between promoters and detractors is your NPS. High NPS indicates product is good enough to spread organically. Low NPS means you need to fix product before investing in growth.

Actual referral rate matters more than intent. NPS measures what humans say they will do. Referral rate measures what they actually do. Saying and doing are different things. Track what percentage of users actually invite others. Then track what percentage of invited users activate. This reveals if referral program has meaningful impact or just makes you feel good.

Part 3: Learning Metrics for Experiments

Learning metrics measure whether experiments teach you anything. This is where most humans fail completely. They run experiments but learn nothing because they measure wrong things or measure right things wrong way.

Experiment Velocity

Experiment velocity is how many valid tests you complete per month. Not how many tests you start. How many you complete and learn from. Most companies run one or two meaningful experiments per quarter. Winners run ten to twenty per month.

But velocity without validity is waste. Better to run one experiment correctly than ten experiments incorrectly. Valid experiment requires clear hypothesis, proper sample size, controlled variables, and statistical significance. Most humans skip these requirements because they slow down velocity. Then they wonder why experiments do not teach them anything.

When following A/B testing frameworks for B2B SaaS, remember that small tests teach you about tactics. Big tests teach you about strategy. Button color test tells you which color converts better. Pricing model test tells you how customers value your product. Most humans waste time on small tests because they are safer politically. Winners take bigger risks on tests that matter more.

Statistical Significance and Sample Size

Statistical significance tells you if result is real or random noise. 95% confidence level is standard. This means if you ran experiment 100 times, result would be same 95 times. Most humans declare victory after seeing any positive movement. This is mistake that leads to false conclusions.

Sample size determines how long experiment must run. Small difference requires large sample. Large difference requires small sample. If you need 100,000 users to detect 1% improvement, you do not have 100,000 users, you cannot validate this test. Most humans start experiments they cannot finish. Better to test bigger changes that require smaller samples.

Duration matters as much as sample size. Running test for one day captures daily patterns but misses weekly patterns. Running test for one week misses monthly patterns. SaaS with monthly billing cycle should run experiments at least one full billing cycle. Otherwise you measure signup behavior but miss retention impact.

Learning Quality Metrics

Humans track wins and losses. Did experiment succeed or fail? This is wrong question. Right question is what did you learn that changes your strategy?

Failed experiment that teaches you truth about customer behavior is more valuable than successful experiment that teaches you nothing. Amazon tested wrong things and got wrong answer. Netflix used data as input but made human decision. Difference was not in data. Difference was in courage to learn beyond what data could prove. This is lesson from Rule 64 - being too rational or too data-driven can only get you so far.

Track how many experiments change your product roadmap. If you run 20 experiments and none of them change what you build, you are testing wrong things. Experiments should generate insights that redirect resources. Most humans run experiments to validate what they already decided to do. This is not learning. This is confirmation bias.

Distinguish between big bets and small bets. Small bets test tactics. Button colors. Email subject lines. Below-fold optimizations. These tests produce incremental improvements. Big bets test strategy. Pricing models. Core features. Target customers. These tests produce step-change improvements or complete failures. You need both types but must understand difference. Learning from rapid experimentation marketing shows that most companies default to small bets because they are politically safe, even though big bets create more value.

Measurement Quality

Your metrics are only as good as your measurement systems. Garbage in, garbage out. Most SaaS companies have broken analytics that produce false confidence.

Attribution accuracy determines if you understand what drives growth. Did customer find you through SEO, paid social, email, or referral? In reality, customer probably touched multiple channels. They saw ad. Googled your brand. Read review. Clicked email. Signed up. Which channel gets credit? Most companies give all credit to last touch. This systematically undervalues awareness channels and overvalues conversion channels.

Tracking coverage reveals blind spots. What percentage of user journey can you actually measure? Privacy filters, ad blockers, multiple devices, and dark social create massive gaps in data. Customer discusses your product in Discord. Texts friend. Mentions in Slack channel. None of this appears in dashboard. Then they click Facebook ad and you think Facebook brought them. Understanding these limitations is critical when implementing data-driven SaaS marketing strategies.

Data freshness affects decision quality. Real-time dashboard helps you catch problems early. Week-old data means problems compound before you notice. Most SaaS companies look at metrics once per week in Monday meeting. By the time they notice trend, it is already two weeks old. Winners check key metrics daily and investigate anomalies immediately.

Part 4: What Not to Track

Knowing what to ignore is as important as knowing what to track. Humans love collecting metrics. More dashboards. More charts. More data. But more metrics do not create more clarity. They create more confusion.

Vanity metrics feel good but teach nothing. Total users, total pageviews, total downloads. These numbers always go up if business is alive. They go up even when business is dying slowly. Much better to track active users, engaged users, paying users. These metrics can go down, which makes them useful.

Blended metrics hide important patterns. Average customer lifetime value across all segments masks that enterprise customers have 10x LTV of small business customers. Segment everything. By customer size. By acquisition channel. By signup date. By feature usage. Patterns appear in segments that disappear in averages.

Lagging indicators tell you what already happened. Revenue is lagging indicator. It measures decisions customers made weeks or months ago. Leading indicators predict future. Engagement frequency is leading indicator of retention. Trial activation rate is leading indicator of conversion. Focus on metrics that let you intervene before bad outcomes happen.

Metrics you cannot influence waste attention. Humans track metrics because they are interesting, not because they can improve them. Every metric in your dashboard should connect to action you can take. If metric goes up, what do you do? If it goes down, what do you do? If answer is nothing, remove metric from dashboard.

Part 5: Building Measurement Framework

Framework organizes metrics into hierarchy. Not all metrics are equal. Some predict success. Some measure success. Some distract from success.

Tier one is your north star metric. Single number that defines if you are winning. Every experiment should ultimately move this metric. If experiment does not connect to north star, question why you are running it.

Tier two is input metrics that drive north star. Acquisition, activation, retention, revenue, referral. Focus on bottleneck that limits growth most. Early stage focuses on activation. Growth stage focuses on acquisition. Mature stage focuses on retention and expansion. Do not try to optimize everything simultaneously.

Tier three is diagnostic metrics that explain why input metrics move. When activation rate drops, diagnostic metrics tell you where users abandon. Is it during signup? During onboarding? During first use? Diagnostic metrics guide where to experiment next.

Create simple dashboard with three sections. Top shows north star metric and whether it trends up or down. Middle shows input metrics and which ones limit north star most. Bottom shows recent experiments and what you learned. This dashboard should fit on one screen. If you need multiple screens, you are tracking too much.

Weekly review process keeps team aligned. Monday meeting covers what happened last week and what experiments to run this week. Not what everyone is busy doing. Not vague discussion of strategy. Specific review of metrics, experiments, and learning. This connects to implementing growth marketing dashboards that drive action rather than just display information.

Conclusion

Humans, game is clear on this point. Metrics you track determine experiments you run. Experiments you run determine what you learn. What you learn determines how you win.

Most humans track wrong metrics. They optimize for vanity numbers that make them feel productive. They run small experiments that produce small results. They celebrate meaningless wins and miss valuable failures. This is why they lose slowly while feeling busy.

Winners pick one north star metric that captures value delivery. They identify input metrics that drive north star. They run experiments with proper statistical rigor. They learn from failures as much as successes. They build measurement systems that reveal truth rather than confirm bias.

You now understand what metrics should I track in SaaS growth experiments. Knowledge without action is worthless. Choose your north star. Identify your bottleneck input metric. Design experiment to move it. Measure properly. Learn what works. This is how game is played.

Most humans will ignore this framework. They will continue tracking everything and learning nothing. But you are not most humans. You understand that right metrics reveal path to victory. Wrong metrics hide path while making you feel productive.

Game rewards those who measure what matters and ignore what does not. Your odds just improved. Use this advantage.