Most retention reports are unreadable because the team never decided what “using the product” means. That is a product question, not an analytics one.

What Events Should I Use for Retention?

The birth event should mark the start of a user’s relationship: Account Created or Trial Started. The returning event should be your core value action — the thing that makes the product worth paying for. For a project tool that is Task Completed; for a writing app, Document Edited.

Using Session Start or a login event as the returning event produces flat, flattering curves. Logging in is not value. Every product I have seen with a suspiciously high retention number was counting logins.

How Do I Read the Retention Curve?

Three readings matter, in order.

The shape beats the number. A curve that drops to 20% by day 7 and holds flat at 18% by week 12 is healthier than one at 35% by day 7 that decays to 4% by week 12. The flat tail is the durable user base.

Should I Use N-Day, Unbounded, or Week-Based Retention?

Mixpanel offers all three. N-day counts a user as retained only if they return on exactly day N. Unbounded counts them if they return on or after day N. Week-based buckets the same way over seven-day windows.

Unbounded looks better and is less useful for products with weekly usage patterns — email or invoicing tools get inflated scores. Use N-day for daily-use products and week-based for weekly ones. Then keep it consistent, because switching mid-quarter makes every comparison invalid.

How Do I Split Cohorts to Find What Actually Works?

Break the report down by a property captured at signup or at first use. The useful splits are behaviour, not demographics:

• Did the user complete setup in the first session?
• Which acquisition channel brought them?
• Did they invite a teammate in week one?
• Which plan or trial type did they choose?

Invite-a-teammate splits tend to show the biggest gap in B2B products — often 3-5x week-8 retention. That finding usually survives the correlation-versus-causation objection better than most, though forcing invites is not guaranteed to work.

How Do I Use Cohorts Without Getting Fooled?

Compare like windows. A cohort created three weeks ago has less data and always looks better at week 4 than a cohort from nine months ago — survivorship hides in the shorter window.

Also watch mix effects. If you changed pricing in March, every cohort after March reflects the change, so a product comparison across that line measures price, not product.

Finally, the free plan keeps 90 days of history. Cohort analysis is the report most damaged by that limit — a 12-week retention curve barely fits, and year-over-year comparison is impossible without upgrading.

When Is Retention Analysis a Distraction?

Before you have 500 users. At 50 users per cohort, the noise is larger than any trend you will detect, and three bad signups move the curve by six points. Ship, instrument, and revisit retention once the denominators are above a few hundred.

Quick Summary

• Retention needs a birth event (signup) and a core-value returning event, not a login.
• The flat tail of the curve matters more than the day-7 number.
• Use N-day for daily products, week-based for weekly ones, and never switch mid-quarter.
• Behaviour splits (setup completed, teammate invited) outrank demographic ones.
• Below 500 users per cohort, the noise exceeds the signal.

The SDK install takes ten minutes. Getting event data you can trust for a year takes a taxonomy document and the discipline to enforce it.

What Is Mixpanel and What Does It Answer?

Mixpanel is product analytics: it answers “what do users do inside the product” rather than “where did traffic come from.” Traffic questions belong in Google Analytics; behaviour questions belong here.

A useful distinction: GA tells you 4,000 people landed on the pricing page. Mixpanel tells you 300 of them started a trial and 40 reached the second onboarding step.

How Do I Install Mixpanel and Identify Users?

For web, the snippet goes in the head and initializes with your project token.

mixpanel.init('YOUR_PROJECT_TOKEN', { track_pageview: true })
mixpanel.identify(user.id)
mixpanel.people.set({ plan: 'pro', signup_date: '2026-10-22' })

Call identify immediately after login and on every page load for a logged-in user. Skipping it means every session becomes an anonymous distinct ID, funnels show inflated step drop-off, and you spend a week debugging data that was never wrong — just unattributed.

Anonymous-to-identified merging works if you call identify within the same browser session. Cross-device merging requires you to use your own user ID, not Mixpanel’s generated one.

How Should I Name Events?

One convention, enforced in review. The most common working pattern is Object Action in past tense: Onboarding Completed, Report Exported, Trial Started.

Rules that prevent the mess:

• Never put data in the event name — Report Exported PDF becomes a thousand events. Use a property.
• Keep registered properties under 30. Mixpanel flags high-cardinality ones and they slow queries.
• Track the same event name across web and mobile so funnels span surfaces.
• Reserve $-prefixed events for Mixpanel’s automatic ones; do not create your own.

How Do I Build a Funnel That Answers Something?

Go to Reports → Funnels, pick three to five events in order, and set a conversion window. A signup funnel normally uses 7 days; a checkout funnel uses 1 hour.

A funnel with eight steps tells you nothing actionable. If you want the whole onboarding flow, split it into two funnels and look at them separately.

Break down by a property that could explain the drop — plan type, acquisition channel, browser. The breakdown is where the finding usually is; the top-line percentage rarely is.

What Does Mixpanel Cost?

The free plan includes 20 million events per month, which is far more generous than most competitors, plus unlimited seats and 90 days of data history. The 90-day window is the real constraint: cohorts older than that disappear unless you pay.

Growth pricing is usage-based and starts around $28 per month at low volume, scaling with monthly tracked users. Watch MTU billing — a marketing site with high anonymous traffic can push you into a tier you did not plan for.

Where Mixpanel Gets Annoying

Event volume costs go up with instrumentation, so teams either over-track and pay, or under-track and guess. There is no cheap middle until you have a strong opinion about which 30 events matter.

Data retention on the free plan is also a real limit for annual cohort analysis. Teams that sign up, instrument for six months, and then want year-over-year retention discover the gap at the worst time.

Who Should Skip It

Pre-launch products with under 100 users. There is not yet a behaviour pattern to analyze, and the instrumentation work is better spent shipping. Also skip it if your questions are all revenue or marketing attribution — that is a different tool category.

Quick Summary

• Mixpanel tracks in-product events; use GA for traffic, Mixpanel for behaviour.
• Call identify on every logged-in page load or funnels will be wrong.
• Name events Object Action and keep varying data in properties.
• Free tier: 20M events/month, unlimited seats, 90-day history.
• Keep funnels to five steps and always break down by a property.