I keep seeing the same pattern in mobile products.
Teams set retention as a primary success metric.
Dashboards glow green.
DAU/MAU improves.
And somehow, over time, the UX starts feeling… harder to leave.
Not because anyone set out to manipulate users — but because metrics don’t just measure behavior, they shape it.
Retention doesn’t distinguish value from friction
Retention metrics are blunt instruments.
They don’t tell you why users return — only that they do.
A user who comes back because the product solved a real problem
looks exactly the same as a user who comes back because:
notifications won’t stop
logout is buried
leaving triggers multiple interruptions
From the metric’s perspective, both are success.
From the user’s perspective, they’re very different experiences.
How dark patterns emerge without bad intent
Most dark patterns aren’t the result of malicious design meetings.
They emerge naturally from optimization pressure.
If retention is the goal, the easiest wins tend to look like:
making exit flows harder to find
adding “Are you sure?” friction when leaving
nudging users back “just in case”
defaulting to opt-in rather than opt-out
Each decision is defensible on its own.
Together, they add up to a product that keeps users inside longer — whether it’s good for them or not.
The dashboard improves.
User autonomy quietly erodes.
Retention vs. real engagement
Retention rewards presence.
Engagement rewards purpose.
A retained user might still be confused, frustrated, or trying to leave.
A genuinely engaged user comes back because there’s value.
When teams optimize for retention alone, they often stop asking:
Did the user actually finish what they came for?
Could they leave easily once they did?
Are we earning their return — or engineering it?
That’s the line where optimization turns into manipulation.
The psychological cost of “sticky” design
Dark patterns work because they exploit normal human behavior:
loss aversion
interruption sensitivity
habit formation
Users often feel like they’re choosing to stay — even when the interface is steering them.
Over time, this creates a quiet trust problem:
people don’t feel helped, they feel managed.
And once users notice that, retention usually collapses anyway.
What changes if you design for exit
If you design with exit in mind, metrics start to look different.
You might measure:
task completion without prompts
clean exits after success
clarity of opt-out paths
voluntary return after time away
These metrics are harder to optimize — but they align better with user value.
Retention stops being the goal.
It becomes a side effect.
Metrics are never neutral
Retention metrics feel objective, but they encode assumptions:
that more time is better
that return equals value
that staying is success
Once those assumptions go unquestioned, dark patterns stop being exceptions and start becoming normal UX.
The uncomfortable truth is this:
if you reward retention above all else, you shouldn’t be surprised when products optimize for keeping users in — not helping them out.
Closing note
I’ve been writing more about product incentives, user trust, and design decisions like this recently.
The longer version of this piece — and related essays — live on my site.

Top comments (1)
Some comments may only be visible to logged-in visitors. Sign in to view all comments.