Retention-driven growth

Measure the right thing

Most teams measure "retention" and mean six different things. Pin down three specific metrics:

1. Day-1, Day-7, Day-30 retention

Of users who signed up on day N, what percentage were active 1, 7, and 30 days later? This is the cleanest definition. Use cohort analysis — group users by signup week, then track each cohort's retention over time.

2. Activation rate

What percentage of signups complete the "aha" event — the thing that makes them understand the product? Examples:

• Social app: made their first post, or followed 3 people
• Productivity: completed their first task
• Fitness: completed their first workout
• E-commerce: made their first purchase, or added to cart

Activation rate is the upstream predictor of retention. Users who activate retain at 3-5x the rate of users who don't. If your activation rate is under 50%, everything downstream is struggling to compensate for an onboarding problem.

3. Engaged users / total users

Of all your signed-up users, what percentage were active in the last 7 days? This is the health of your whole base, not just new cohorts. Growing total users while this number shrinks means you're adding users faster than you're keeping them.

If you have the Layers SDK installed for attribution, you already have the event stream you need for these. Fire a signup event, an activation event (your choice), and an app_open event on launch. Everything else falls out of those three.

Find the leak

Look at your signup funnel step by step. Where do people drop off?

A typical mobile app funnel:

100%  Install
 85%  First app open
 55%  Completed onboarding
 35%  Key action performed (activation)
 18%  Active on day 2
 12%  Active on day 7
  6%  Active on day 30

The biggest drop in that example is between "completed onboarding" and "key action performed" — 35% to 18%, almost half of activated users don't come back on day 2.

For most products, the biggest leaks are one of four places:

1. Install → first open. If 15%+ of installs never open, you have a metadata or expectation problem. Users installed based on a promise that the app didn't match in the first 5 seconds. Usually a creative/brand mismatch, or an onboarding wall users bounced off.

2. First open → onboarding complete. If less than 65% of openers finish onboarding, your onboarding is too long or too confusing. Count the screens. Cut two.

3. Onboarding complete → key action. If less than 50% of onboarded users hit your key action within 24 hours, onboarding technically completed but didn't actually activate the user. You onboarded them to a menu, not to doing the thing.

4. Day 1 → Day 2. If more than half of day-1 users don't return on day 2, they didn't have a reason to come back. No notification, no hook, no reason. This is where retention loops matter.

You can have leaks at all four places; they compound multiplicatively. A 50% leak at each step means 6% retention at the end. Fixing one step to 75% gets you to 14% — more than double, from a single fix.

What failure looks like

You look at your retention chart, see it's bad, and decide to "do a retention push." You send a flurry of notifications, launch a referral program, and build a "week 1 engagement streak" feature. None of it moves the number because the actual leak was in step 2 of onboarding, where users hit a confusing permissions screen. Fix: find the specific leak step before building any retention feature.

The first retention fix: shorten onboarding

For 70% of consumer apps, the fastest retention improvement comes from cutting onboarding length. Count your onboarding screens right now. Whatever the number, cut it in half.

Yes, half. Yes, you'll lose information you think is critical. Users don't need that information to get started. They'll figure out the advanced stuff once they're hooked.

Onboarding rules that work:

• No more than 3 screens before the user does something real. One intro, one value prop, one permission/setup, then straight to the product.
• Defer permission requests until the moment they're needed. "Allow notifications?" on screen 2 gets denied. "Allow notifications so we can tell you when your workout is ready?" on screen 9, after they've set up a workout, gets accepted.
• No feature tutorials. Show, don't tell. Let them explore and discover.
• One clear CTA per screen. If there are three buttons, people pick none.

Most apps double their onboarding-complete-to-activation rate just from cutting onboarding length. This is the single highest-leverage retention intervention available to most teams.

The second retention fix: day-2 hook

If users don't come back on day 2, nothing else matters. Your job is to give them a reason.

Options that work:

Notification hooks. One notification on day 1 evening or day 2 morning, referencing something specific the user did. Not "come back!" — something like "You started a plan — ready for day 2?"

Email hooks (if applicable). Same principle, different channel. Works better for utility apps, worse for social/gaming.

Feature completion loops. Design one feature that has a multi-day cadence — a streak, a scheduled session, a "check back tomorrow" rhythm. This pulls users back naturally.

Social feeds. If there's a social component, populated feeds pull users back. Empty feeds don't.

Options that rarely work:

• Generic "we miss you!" re-engagement pushes
• Long welcome email sequences
• "Tip of the day" broadcasts

The common thread: what works is specific to what the user already did. What doesn't work is generic broadcast.

The third retention fix: week-1 compound value

For users who come back on day 2, the next question is whether they come back on day 7. Retention past the first week usually means the product has started compounding for them.

Compounding value comes in three flavors:

1. Investment. The user has put something into the product that's worth coming back to. Data, customizations, progress, a streak, content they created, connections they made. Dating apps work because of the profile investment. Productivity apps work because of the task history. Gaming works because of progress.

2. Habit. The user has built a routine that includes your app. Notification at the right time, placement on their home screen, muscle memory. This takes weeks to build; you can only help it along.

3. Social gravity. Their friends are on the app, or their friends will be if they keep using it. This is the strongest form but the hardest to manufacture.

Your product needs at least one of these three to retain past week 1. If users finish your onboarding and there's no investment, no habit formation, and no social layer, they will leave — not because you did anything wrong, but because there's no reason to stay.

This is not a marketing fix. It's a product fix.

When retention is "fixed enough" to restart acquisition

You don't need perfect retention to turn acquisition back on. You need retention that's proven enough that acquisition spend pays back.

Specifically:

• Activation rate above 40% of signups
• Day 7 retention within range of your vertical's benchmark (doesn't have to match; has to be in the same order of magnitude)
• Day 30 retention on an upward trend — even if it's not at benchmark yet, it's improving cohort over cohort
• LTV > CAC with reasonable assumptions for your target CAC

Once these are true, acquisition starts working. Before they're true, acquisition wastes money no matter how well-targeted.

This is why first ads on Meta and Apple Search has retention as a prerequisite. Without it, the ad spend optimizes toward CAC, you feel like you're winning, and the users churn out the back. A year later you've spent a lot on ads and have the same DAU you started with.