In-App Loop

The Loop Itself

Where engagement actually happens

The loop inside the app

The Digital Value Loop describes a user’s whole journey on the platform, but most of the meaningful engagement happens inside the app, after a user has installed and registered. The In-App Loop is the part of the cycle where the day-to-day relationship between user and product gets built, tested, and either reinforced or broken.

This section breaks the in-app journey into five stages, each with its own dynamics and its own questions to answer:

1. The First Day: what does the user need to commit to so they come back tomorrow?
2. Short-term Retention: what brings them back in their first weeks?
3. Long-term Retention: what keeps them around for months?
4. Re-engagement: how do we bring them back once they’ve stopped?
5. Churn: when do we accept the user has moved on, and what can we learn?

Each stage has different drivers

A common mistake is to treat retention as a single number to optimise. In reality, the levers that bring a user back on Day 1 are not the same as the ones that keep them around in Month 6.

On the first day, commitment-building actions (setting a goal, sharing data) do most of the work. By six months in, brand affiliation and product-market fit with a specific persona matter much more. Use the wrong lever at the wrong stage and you’ll waste effort.

Why split it this way

Treating the loop as five distinct stages makes it possible to:

• Diagnose where users are dropping out, and where they aren’t
• Match interventions to the actual problem at each stage
• Measure each transition separately rather than collapsing everything into one retention number

The sections that follow walk through each stage with the patterns and data points learned from a real digital product over four years.

The First Day

The single most important session

Why the first day matters more than any other

The data is clear: what a user does in their very first session has an outsized impact on whether they ever come back. Users who complete certain commitment actions in their first day stay around at multiple times the rate of those who don’t. This is the highest-leverage moment in the whole engagement loop.

Commitment equals retention

Whether through sunk-cost fallacy or positive reinforcement, a user who feels they’ve committed to some future outcome is much more likely to come back for a second day.

We call this “Activation”: completing in-session actions that anchor the user to a future they care about.

Examples of activation actions, drawn from the LiveWell journey:

• Selecting a specific health goal
• Sharing personal health information such as habits or weight
• Connecting a health data source like Apple Health
• Accepting marketing or push notifications

The numbers

In the LiveWell data set, activated users were 4x more likely to be active in their second week than non-activated users (38% retained vs. 9%). This held across cohorts and was the single strongest first-day predictor we found.

Skipping onboarding is a leading indicator of churn

Users who skip onboarding don’t just have lower retention. They also complete fewer in-app activities while they are active. Skipping is a signal that the user has not yet decided whether the product is for them, and they tend to drift away soon after.

What this means for design

The first session should be designed to extract one or two meaningful commitments early. Not by force, but by making the value of committing obvious to the user. A vague onboarding flow that asks for nothing in return feels polite, but it quietly costs you Day 2.

Three things to take away

1. The first session is the highest-leverage moment in the entire loop.
2. Build at least one commitment action into onboarding (a goal, a connection, a consent).
3. Treat skipping the onboarding flow as a strong churn signal, and design with that in mind.

Short-term Retention

Getting users to come back within their first month

What we mean by short-term

Short-term retention covers the period from a user’s second visit through to roughly their first month. The first day is about commitment. This stage is about proving the product is worth coming back to.

Always buy the first coffee

Engagement at this stage looks a lot like a game of tit-for-tat. Users are far more likely to engage again if they got something useful for free in their first session. A/B testing on LiveWell showed that offering a free downloadable (a guide, a plan, a checklist) lowered initial skepticism and made future engagement noticeably easier.

The principle generalises: give first, then ask. The “ask” can be a notification opt-in, a profile completion, or a paid upgrade later, but the value has to flow toward the user first.

Effort at the start pays off

Counter-intuitively, the users who stick around are usually the ones who attempted higher-effort activities on Day 1, not the ones who only did the easy stuff. In LiveWell data, returning users were many times more likely than non-returning users to have completed activities like:

• Tracking 5,000 steps via a connected wearable
• Reading an article
• Joining a Boost (community event)
• Completing a meditation session

These activities work because they double as a tutorial. The user doesn’t just see what the app can do, they experience it.

A small nudge goes a long way

For users who didn’t come back on their own, a single well-timed push notification 48 hours after enrolment recovered a meaningful share of them. In one LiveWell cohort of around 47,700 users, roughly 19% visited the app the same day they received this notification.

The lesson: don’t assume a non-returning user has churned permanently. The first two days of silence are often a memory problem, not a value problem, and the cost of a single well-targeted nudge is very low.

Three things to take away

1. Offer something of perceived value in the first session, before asking for anything back.
2. Design the first session so the user performs an action that demonstrates what the app actually does.
3. A timely, low-friction reminder in the first 48 hours is worth setting up.

Long-term Retention

Users who engage for multiple months

What we mean by long-term

Long-term retention is users active beyond roughly six months. By this point, the user has chosen to keep the app in their life. The drivers here look very different from the drivers in the first month.

Affiliation is the strongest signal

In LiveWell data, about 75% of users still active after 180 days came through a Zurich BU or partner channel, not through direct download. Brand and partner affiliation isn’t only useful at the top of the funnel; it shows up again here as a predictor of who stays.

The implication: affiliation isn’t only a marketing benefit. It has a long tail through the entire user lifecycle, and acquisition channel quality matters more than acquisition volume.

Product-market fit becomes visible here

By six months in, the users who remain tend to share a clear profile. At LiveWell we call this profile the “wellness seeker”, and 62% of users active past 180 days identify as motivated on their fitness journey when polled. That number is much higher than for the general user base.

This is useful in two ways:

1. It tells you who your real customer is, regardless of who you thought you were building for.
2. It gives you a clear target persona for top-of-funnel campaigns, closing the loop on acquisition.

No single feature keeps everyone engaged

A common temptation is to look for “the one feature” that drives long-term retention. The data doesn’t support that.

Long-term users at LiveWell cluster into different cohorts (meditators, step-trackers, boost-joiners, article-readers), and each cohort leans heavily on a different feature set. The product needs to be broad enough to serve several cohorts, even if no single cohort uses the whole product.

Continuous delivery keeps the product alive

Long-term users notice when a product stops evolving. A short continuous delivery cycle does two things at once:

• It addresses real user issues quickly enough that they don’t accumulate into churn
• It signals that the product is actively cared for, which itself reinforces engagement

Practical implications

For teams working on long-term retention:

• Treat partner and brand affiliation as a retention input, not just an acquisition input.
• Use polls and feedback from your 6-month-plus cohort to clarify your real persona.
• Build for several user cohorts, not one composite “average user”.
• Ship continuously and visibly.

Re-engagement

Bringing previously inactive users back

Re-engagement is its own discipline

Once a user has stopped engaging, the toolkit changes. You no longer have the in-app surface, so re-engagement depends entirely on the channels you set up while the user was still active. Most of the work for re-engagement is done before the user goes quiet.

Understand your reach first

Before any campaign, the basic question is: how many of these users can we actually reach? That means tracking, at minimum:

• Push notification consent rate
• Marketing email consent rate
• Email bounce and unsubscribe rates
• App uninstall signals where available

If only a small share of churned users has given consent to be reached, no campaign will move the needle. Fix consent rates during onboarding and short-term retention before investing heavily in re-engagement.

Timing is everything

Re-engagement has sharply diminishing returns the longer a user has been inactive. iOS, for instance, can begin offloading unused apps after as little as 12 days of non-use, at which point your push notifications quietly stop arriving. By the time you decide a user is “churned” by a 30-day or 90-day rule, your ability to act may already be reduced.

The practical implication: trigger your first re-engagement attempt earlier than feels natural, while you still have the channels.

Live-service events draw users back

Even the largest apps have to work continuously to re-engage their base. At LiveWell, time-limited in-app events called “Boosts” have been the most consistent tool for pulling churned users back, even if only for a short window.

During a January 2025 Boost campaign communicated through Instagram and email to Polar-affiliated users, that segment’s Normalised MAU jumped by roughly 50% before settling back down. The lift is temporary, but the cohort that returns is then available for further engagement, and many will stay longer than the campaign itself.

Leverage partner channels

Partner companies usually have their own established communication channels: social media accounts, newsletters, push capabilities of their own apps. Re-engagement campaigns that ride on these channels reach users in contexts they already trust, and they reach users who may have unsubscribed from your direct comms.

A short checklist

• Are consent rates for push and email being tracked and improved upstream?
• Is there a re-engagement trigger before the user is technically “churned”?
• Is there a calendar of live-service events that can pull lapsed users back in?
• Are partner channels being used in addition to your own?

Churn

When a user has effectively left

Churn is business-dependent

There is no single right answer to “when is a user churned?” For a fitness app it might be 30 days of inactivity. For a tax filing app it would be unreasonable to call a user churned after even 11 months of silence. The definition follows the natural engagement cadence of the product, not a generic rule.

A useful starting point:

• Daily-use products (social, messaging, news): churn at roughly 7 to 14 days of inactivity
• Weekly-use products (fitness, learning, productivity): churn at roughly 30 to 60 days
• Monthly or seasonal products (tax, travel, holiday planning): churn measured in quarters or longer

Whatever threshold you pick, be explicit about it and use it consistently across the team.

Churn is not the same as abandonment

It’s worth keeping two related ideas separate:

• Abandonment happens during onboarding, before the user ever entered the engagement loop. The dashed lines in the Digital Value Loop diagram show this: failed downloads, incomplete registration, drop-off before the first meaningful session.
• Churn happens after the user has entered the loop. They engaged, and then they stopped.

The distinction matters because the response is different. Abandonment is an onboarding problem (build intent earlier, simplify the flow). Churn is an experience or fit problem (the product stopped delivering, or never matched what the user actually wanted in the first place).

Not all churned users are gone

A user not engaging this month does not mean a user who will never come back. As long as a contact channel is still open (push consent, email, presence on partner platforms), there is a real chance to bring them back. See the Re-engagement section above for the levers that work.

The implication is that what you do during the active period determines what’s possible during the churned period. If consent and channel set-up were neglected earlier, churn is final by default.

Use churn as a learning signal

Churn is also one of the highest-signal data points the product has. A few useful questions to ask of every churn cohort:

• What was the user’s last action before going quiet?
• Did they reach activation? If not, what did they skip?
• Which cohort do they sit in (meditator, step-tracker, browser of articles, etc.)?
• Did they churn alone, or as part of a wider drop (a bug, a botched release, a campaign that brought in poorly-matched users)?

The answers feed directly back into the earlier stages of the loop.

Bottom line

Treat churn as an expected stage of the loop, not a failure state. Define it deliberately. Distinguish it from abandonment. And design the active stages so that, when churn does come, the doors back in are still open.