I downloaded a news app recently.
Opened it. Scrolled a bit. Then I saw a section labelled:
“For You.”
Naturally, I expected something… relevant.
Instead, I saw a list of stories I had zero interest in.
Not slightly off. Completely off.
At first, it felt like a small annoyance. Then it became something more interesting.
The product was confident enough to say this is for you while clearly not understanding me at all.
That’s where the irony of personalisation starts.
When “personal” doesn’t feel personal
Personalisation is supposed to make products feel tailored.
More relevant content.
Less noise.
Better decisions made on your behalf.
But when it misses, it misses in a very specific way.
It doesn’t just feel irrelevant.
It feels wrong.
A generic feed is easy to ignore. You expect randomness.
A personalised feed carries an assumption. It claims to know you.
So when it gets things wrong, it breaks trust faster.
The problem with early assumptions
Most personalisation systems don’t actually know the user yet.
They rely on:
- limited initial signals
- general trends
- location data
- popular content
In the case of Google News, the “For You” section appears almost immediately.
The issue is timing.
The product starts making decisions before it has enough information.
Instead of feeling personalised, it feels like guesswork presented as certainty.
Confidence without accuracy
There’s something subtle happening here.
The label “For You” is not neutral. It’s a strong claim.
It tells the user:
“We understand your preferences.”
But the system hasn’t earned that confidence yet.
This creates a mismatch between:
- What the product says
- what the user experiences
That gap is where trust starts to drop.
Why this matters more than it seems
Personalisation is not just about relevance. It’s about credibility.
If a product consistently shows content that feels off, users start to question it.
They stop relying on it.
They ignore recommendations.
They mentally downgrade the product.
In some cases, they disengage entirely.
It’s not just a content problem. It’s a perception problem.
A better way to think about personalisation
The issue is not personalisation itself. It’s how it is introduced.
Instead of acting confident too early, products could:
- start with neutral language
- signal that they are still learning
- allow users to shape the experience quickly
- gradually earn the right to personalise
There’s a difference between:
“For You”
and
“We’re learning what you like”
One assumes. The other builds trust.
When personalisation actually works
Good personalisation feels quiet.
It doesn’t announce itself loudly.
You just notice that things start to feel more relevant over time.
It adapts.
It improves.
It reflects your behaviour.
The key difference is that it earns its accuracy instead of assuming it.
Final thought
Personalisation works best when it is felt, not declared.
The moment a product claims to know you, it raises expectations.
If it cannot meet those expectations yet, the experience feels worse than no personalisation at all.
Sometimes it’s better to be honest and generic at the start than confidently wrong.
Because nothing feels less personal than being told something is “for you” when it clearly isn’t.
