There is no single most important metric for every startup. The right metric depends on the business model, stage and the behaviour that creates value. A marketplace, SaaS project, media site and e-commerce store should not use the same dashboard.
Visits, users, sessions and hits: what do they mean?
Visits
A general term for visits to a site. Different analytics tools may define it differently.
Users
An estimate of distinct people or devices. Cookie restrictions and cross-device behaviour make this an approximation.
Sessions
A group of interactions during one visit period.
Hits
An older technical term for requests to a server. It should not be used as a proxy for people because one page can generate many hits.
Pageviews
The number of pages viewed, including repeated views by the same person.
DAU, WAU and MAU: how many people actually use the project?
DAU - daily active users
Users who perform a defined meaningful action in a day.
WAU - weekly active users
Useful for products with a weekly rhythm.
MAU - monthly active users
Useful for broader reach but can hide weak frequency.
DAU/MAU ratio or stickiness
Shows how frequently monthly users return. The meaning depends heavily on the product: a tax tool should not be judged like a messaging app.
Acquisition: where do people come from?
Track channels such as direct, organic search, paid search, social, referrals, partners, email and sales. Use consistent campaign parameters so you can connect traffic to downstream behaviour.
Funnel: how many people reach the sale?
A funnel converts one stage into another: visitors → signups → activated users → trial users → paying customers. Measure conversion between each step so you know where the biggest loss occurs.
About the example “5% sign up, 2% become paying customers”
Always define the denominator. Two percent of all visitors and two percent of signups are very different numbers.
What is a good conversion rate?
There is no universal benchmark. Intent, price, channel, device, region and category all matter. Your own historical baseline and the economics of the funnel are often more useful than generic industry averages.
Activation: did the user experience the value?
Activation should represent an “aha” moment, not merely account creation. For a design tool it might be exporting the first design; for a marketplace, completing the first transaction; for an education app, finishing the first lesson successfully.
Retention: do people come back?
Retention measures whether users who started in a period remain active later. Cohort analysis is useful because it separates new users from older ones.
A stabilising retention curve is one of the strongest signals that some users continue receiving value.
Churn: how many customers leave?
Customer churn
The percentage of customers lost in a period.
Revenue churn
The percentage of recurring revenue lost.
Gross revenue retention (GRR)
How much starting recurring revenue remains after cancellations and downgrades, excluding expansion.
Net revenue retention (NRR)
Includes expansion from existing customers. NRR above 100% means the existing customer base grows even before new customers are added.
MRR and ARR: recurring revenue
MRR - monthly recurring revenue
Normalised recurring subscription revenue per month.
ARR - annual recurring revenue
Often approximated as MRR × 12 for stable subscription businesses, but one-time fees should not be included as recurring revenue.
Growth rate: how fast is the company growing?
Growth can be measured month over month or year over year. Small companies can show very high percentages from a tiny base, so always look at the absolute numbers too.
CAC: how much does a customer cost to acquire?
Customer acquisition cost should ideally include relevant sales and marketing expenses, not only advertising spend. If €10,000 of acquisition cost produces 100 new customers, CAC is €100.
LTV and the CAC relationship
Lifetime value estimates the economic value of a customer over the relationship. LTV/CAC is useful only if both numbers are calculated realistically. Inflated lifetime assumptions can make an unprofitable channel look healthy.
ROI and ROAS
ROAS compares advertising revenue with ad spend. ROI compares profit with the broader investment. A campaign can have positive ROAS and still lose money after product costs and overhead.
Burn rate and runway: how long can the company survive?
Net burn
Cash outflow minus cash inflow over a period.
Runway
Cash balance divided by net monthly burn. If a company has €300,000 and burns €30,000 per month, the simple runway is about ten months.
Burn multiple
Burn multiple compares net burn with net new recurring revenue and is one way investors evaluate capital efficiency in SaaS. Like all metrics, it is most useful in context rather than as a universal pass/fail score.
Rule of 40 and investor perspective
The Rule of 40 is a common heuristic for more mature software companies: growth rate plus profit margin around 40% is often considered a strong balance. It is not a sensible target for every early startup.
What might investors consider “good”?
Expectations vary by stage, market and year. Investors typically look for combinations of growth, retention, efficient acquisition, gross margin and a credible market opportunity rather than one magic metric.
Industry examples: what should different projects watch?
SaaS
MRR, ARR, activation, churn, NRR, CAC, payback and gross margin.
E-commerce
Conversion, average order value, gross margin, repeat purchase rate, return rate and blended CAC.
Marketplace
Gross merchandise value, take rate, liquidity, repeat transactions, supply quality and contribution margin.
Consumer or mobile app
Activation, DAU/MAU, cohort retention, session frequency, organic sharing and monetisation.
B2B sales
Qualified pipeline, win rate, sales-cycle length, contract value, expansion and retention.
Content or media
Reach, return visits, engaged time, email subscribers, paid conversion, ad yield and direct audience share.
Vanity metrics
Likes, impressions, downloads and registrations can be useful diagnostics, but they become vanity metrics when they are presented as success without a link to value or business outcomes.
North Star Metric
A North Star Metric is one primary measure that represents value delivered to customers. It should be specific to the project. “Revenue” is important, but a behavioural metric can sometimes reveal product health earlier.
How to measure: Google Analytics, product events and Counter.dev
Google Analytics 4
Useful for traffic sources, pages and web events, though privacy and attribution limitations should be understood.
Counter.dev
A lightweight privacy-focused option for simple website analytics.
The project database as the source of truth
For signups, payments, subscriptions, completed actions and retention, your own database often provides the most reliable operational record.
What simple dashboard does an early startup need?
Keep it small: visitors or leads, activation, weekly or monthly active users, paid customers, revenue, retention/churn, acquisition source and cash runway. Add metrics only when they change decisions.
The most important thing is not a “good number” but knowing why it changed
Metrics become useful when they connect to behaviour. If conversion falls, ask which source, segment or step changed. If retention improves, find out what those users experienced differently. A dashboard should help you make decisions, not simply decorate investor slides.
Cohort analysis
Aggregate averages can hide important changes. Compare users who started in the same week or month and see whether newer cohorts activate and retain better than older ones. This is one of the clearest ways to see whether product improvements are working.
Segmentation
Break metrics down by customer type, acquisition source, plan, geography or use case. Overall retention may look mediocre while one segment is excellent. That segment can become the focus for product-market fit.
Conversion rate by stage
Do not optimise only visitor-to-purchase conversion. Measure visitor-to-signup, signup-to-activation, activation-to-paid and paid-to-retained. A funnel shows where the real constraint exists.
Sales pipeline metrics
For B2B, track leads, qualified opportunities, demos, proposals, win rate, average contract value and sales-cycle length. A large pipeline is not valuable if deals never advance.
Payback period
CAC payback measures how many months of gross contribution are required to recover acquisition cost. Shorter payback reduces financing risk and makes growth easier to fund.
Gross margin
Revenue is not equally valuable across businesses. Subtract direct delivery costs to understand gross margin. AI inference, cloud infrastructure, support labour, payment fees and physical fulfilment can materially change the economics.
Average revenue per user or account
ARPU or ARPA helps explain whether growth comes from more customers or more value per customer. Combine it with retention so price increases do not hide customer loss.
Expansion revenue
Measure upgrades, additional seats and usage growth from existing customers. Strong expansion can offset churn and produce NRR above 100%.
Logo retention versus revenue retention
You can lose many small customers while keeping most revenue, or lose one large account and keep most customer logos. Track both to understand concentration risk.
Customer concentration
If one customer represents 40% of revenue, the business has a different risk profile from one with hundreds of balanced customers. Concentration is especially important in early B2B companies.
Cash metrics
Bookings, billings, revenue and cash collected are different. Annual contracts can create strong bookings while cash arrives later. Understand what actually reaches the bank account.
Forecast accuracy
As the business matures, compare forecasted and actual revenue, churn and expenses. Improving forecast accuracy is a sign that the system is becoming more predictable.
Qualitative metrics
Not everything important is numeric. Record repeated customer complaints, cancellation reasons, support themes and the jobs users say they would miss. Quantitative data tells you what changed; qualitative evidence often explains why.
Instrumentation quality
Bad event definitions produce false precision. Write down exactly what each metric means, where it comes from and which edge cases are excluded. If two dashboards disagree, resolve the definition rather than choosing the prettier number.
Metric hierarchy
Use a small top-level dashboard for company health and deeper diagnostic metrics underneath. Founders should be able to see the system in minutes and investigate details only when something moves.
Do not optimise metrics in isolation
Higher conversion can reduce retention if the offer attracts the wrong people. Lower support cost can reduce customer satisfaction. Faster sales can increase bad-fit customers. Metrics are connected, so always check the downstream effect.
Activation rate formula
Activation rate is activated users divided by the relevant starting cohort, usually signups. Define activation as a meaningful value event before calculating the percentage.
Retention tables
A cohort retention table shows what percentage of users from each signup period return in later periods. It makes project changes visible: newer cohorts should ideally retain better after important improvements.
Churn and retention are related but not identical
For a fixed customer base, retention is roughly the inverse of churn, but timing and revenue expansion complicate the picture. Define whether you are measuring users, customers, subscriptions or revenue.
MRR movements
Break MRR into new MRR, expansion MRR, contraction MRR, reactivation and churned MRR. This explains how recurring revenue changed instead of showing only the final total.
Quick ratio for SaaS
Some SaaS teams compare new plus expansion revenue with churn plus contraction. The ratio can show whether growth is driven by healthy additions or constantly replacing losses. It should be interpreted together with absolute scale and retention.
Engagement depth
For projects where simple login is not meaningful, measure completed projects, documents created, messages sent, lessons finished or another action that represents real value.
Time to value
Measure how long it takes a new user to reach the first meaningful outcome. Reducing time to value can improve activation and retention without adding features.
Support metrics
Ticket volume, response time and repeated issue categories can reveal project friction. A growing customer base should not require support to grow at exactly the same rate if the project is becoming easier to use.
Reliability metrics
For critical systems, uptime, error rates, latency and failed transactions are project metrics, not only engineering metrics. Reliability directly affects retention and trust.
Experiment metrics
For each experiment, define one primary outcome and guardrail metrics. A change that increases conversion but sharply increases refunds may not be an improvement.