Quantifying engagement and retention in banking apps

Worked example:

A 28.6% stickiness means the average active user opens the app roughly 8 to 9 days per month. For banking, that is healthy. Banking apps run structurally high on stickiness because payments, balance checks, and card management are habitual. Consumer fintech benchmarks commonly cite banking and payments apps in the 40%-plus range for leaders, though this varies widely and figures are estimates that depend on how each firm defines "active." Always check the definition before comparing two banks.

Primary-account share: the metric that pays the bills

Here is the metric marketers in banking obsess over: primary-account share (also called "primary banking relationship" or "main bank" share). It is the percentage of your customers for whom you are the main financial institution.

Why it dominates: a primary customer routes their salary to you, pays bills from you, and swipes your card daily. A secondary customer parks a little money and forgets you.

How to define "primary"

There is no single regulatory definition, so pick observable signals and be consistent:

• Direct deposit / salary inflow present (the single strongest signal).
• Debit card transactions above a threshold (for example, 5-plus per month).
• Recurring bill payments or direct debits set up.

A common working rule: a customer is primary if a recurring salary-type inflow lands AND they make at least a handful of card transactions per month.

Worked example:

This 42% predicts revenue far better than downloads. Neobanks historically struggled here: many customers used them as a spending "side pocket" while keeping their salary at an incumbent. Companies like Monzo and Revolut in Europe, and Chime in the US, have publicly framed growth around converting secondary users into primary ones (driving salary switching), because primary customers generate multiples more interchange and fee activity. Treat any specific share figures from press coverage as company-reported estimates.

For a primer on the digital-banking competitive landscape, the Federal Reserve's payments research hub is a free, credible starting point.

Attrition curves: dormant versus churned

Now the hard part. In a subscription app, churn is obvious: the customer cancels. In banking, almost nobody formally closes an account. They just go quiet. So you must build attrition curves that separate two states.

The two states

• Dormant: the account is open, may hold a small balance, but shows no meaningful activity for a defined period. The relationship is recoverable with the right campaign.
• Churned: the customer has effectively left. Salary redirected elsewhere, balance drained, no logins. Recovery is unlikely.

The trap: treating dormant accounts as churned makes retention look worse than reality and wastes win-back spend. Treating churned accounts as merely dormant flatters your numbers and hides a leak.

Defining the thresholds

Set explicit, time-based rules. A common framework:

| State | Definition (illustrative) |

|-------|---------------------------|

| Active | Meaningful action within 30 days |

| At risk | No transaction in 30 to 89 days, balance still present |

| Dormant | No transaction in 90 to 179 days, balance near zero or minimal |

| Churned | No transaction 180-plus days AND salary inflow stopped |

These windows are illustrative, not regulatory. (Note: some jurisdictions have separate legal definitions of a "dormant account" for unclaimed-funds purposes, for example the UK's dormant assets scheme, but that is a compliance concept, not your marketing metric.) Your marketing definition should be set from your own dataown dataData collected directly from your own customers and prospects through your own channels: your most reliable and privacy-compliant source.View full definition → and applied consistently.

Building the curve

A retention (or survival) curve plots the percentage of a cohort still active over time. Take everyone who opened an account in January, then measure how many remain active at month 1, 2, 3, and so on.

Illustrative cohort of 10,000 new customers:

The curve tells you two things. The steepness of the early drop (Month 1 to 3 here loses 28%) reveals onboarding weakness. And the point where the curve flattens reveals your durable core. If retention stabilizes near 50% after a year, that plateau is your true long-term customer base.

Early-life attrition in banking apps is heavily driven by activation friction: customers who never set up direct deposit or never make a first card transaction churn far faster. That is why the primary-account signal and the retention curve are two views of the same battle.

Putting the three metrics together

These metrics are not independent. They form a funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → of engagement quality:

1. Active-user rate tells you how many people show up.

2. Primary-account share tells you how many treat you as home.

3. Attrition curve tells you how long that lasts.

A bank can have strong MAU but weak primary share (a common neobank pattern: lots of users, few salaries). Or strong primary share but a steep early attrition curve (good product, poor onboarding). Reading all three prevents the single-metric mistake.

A quick diagnostic

Ask these in order:

• Is stickiness (DAU/MAU) above roughly 25%? If not, the app is not habitual.
• Is primary-account share rising quarter over quarter? If not, you are acquiring side-pocket users.
• Does the retention curve flatten, or keep declining past month 12? A curve that never flattens signals a structural leak, not just onboarding friction.

Why the dormant / churned split changes your budget

Marketing spend flows differently by state. At-risk and dormant customers get win-back campaigns (a targeted "set up direct deposit and get X" offer often works because the salary signal is the primary lever). Churned customers rarely justify win-back spend; that budget is better aimed at acquisition or at protecting your active base.

If you misclassify 100,000 churned customers as dormant, you might spend on reactivation offers that almost nobody redeems, then wrongly conclude "win-back doesn't work." The classification error, not the tactic, was the failure.

Key Takeaways

• Downloads and even MAU can mislead. Primary-account share (salary inflow plus regular card activity) is the marketing metric that best predicts real revenue in a banking app.
• Compute stickiness as DAU/MAU. Banking apps are structurally sticky; leaders are often cited above 40%, but always verify each firm's definition of "active" before comparing (figures are estimates).
• Never equate dormant with churned. Set explicit time-and-behavior thresholds (for example, dormant at 90 to 179 days, churned at 180-plus days with salary gone) and apply them consistently.
• Read the retention curve's shape. A steep early drop signals onboarding and activation friction; the plateau reveals your durable customer core.
• Classification drives budget. Win-back spend belongs on at-risk and dormant customers; misclassifying churned users as dormant burns campaign money and produces false conclusions.