Personalization Has a Trust Problem

Personalization was supposed to make marketing more relevant and targeted. Instead, it seems to have made a lot of brands just feel intrusive.

When people talk about “creepy” marketing, they usually talk about the contents of messages or ads they are receiving.

I think that gives the message too much blame. The problem usually starts much earlier, in the logic behind it.

A brand sees a signal, stores it, combines it with other signals, infers something more, then acts on it later in a different context. By the time the customer sees the output, the feeling of being watched is already built in.

I want to buy a blue sailboat

There is a common notion that tech companies, particularly those developing devices that we invite into our homes - our mobile devices, computers, home assistants (think Alexa, Google Home, etc.), and more, are spying on us and taking our information without our awareness - dodging transparency laws and putting us in a weird spot. 

People feel as though they are being watched. And in many cases, they are.

I did an experiment a few years ago to test this out - and I don't know that the results prove anything real - but they reinforce an uncanny narrative. 

I don't know anything about sailing, I don't care about boats, and I barely live near water. 

I decided to start loudly and frequently talking about wanting to buy a sailboat, particularly a blue one, and to take it sailing on the Black Sea - a sea I have never been to and live nowhere near. 

I said this often - not writing, simply speaking - out loud and in the presence of my devices. 

The hypothesis was that maybe the apps on my phone were listening to my conversations to provide personalization. 

Of course, there was certainly a layer of anecdotal paranoia here. I worried that they were, and I created my own self-fulfilling prophecy. 

So when I started receiving advertisements on Facebook and Google for boats and vacation getaways in Bulgaria and Turkey, my fears felt validated. 

Now, were my devices literally listening to me and turning those conversations into ads? There is little public evidence that major platforms like Amazon, Meta, or Google secretly use microphone audio for ad targeting, and these companies deny it.

But that almost makes the story creepier. Ad systems do not need to hear me say ‘blue sailboat on the Black Sea’ to make me feel watched. They can work from other searches, site visits, app activity, location, device identifiers, advertiser data, social signals, and statistical guesses about interests and socio-economic product fit.

So whether the ads were triggered by something I did, something someone near me did, or pure coincidence, the result was the same: the personalization felt invasive enough to damage my trust.

Fair or not, I increasingly find myself less trustful of data collection. Find out why I always reject cookies.

Why personalization started to feel wrong

Most people do not object to relevance, but to relevance that feels unearned, or that comes from somewhere it shouldn't.

There is a big difference between a company responding to what I am doing right now and a company acting like it knows me more intimately than our relationship should allow.

If I browse a product category and then see a related recommendation, that makes sense. If I mention something once, move to another channel, and suddenly encounter a message that feels too specific, the brand has crossed into a different kind of territory.

That distinction matters because personalization is not judged only by accuracy. It is judged by legitimacy. A message can be technically correct and still feel inappropriate.

The real issue is hidden logic

A lot of personalization still runs on invisible chains of collection and inference. Data is gathered and enriched in different, often unknown places. Sometimes it isn't used until long after it is gathered. Then it shows up again in a form the customer cannot easily explain. That is where trust starts to break.

This is why so many brands misunderstand the backlash. They assume the answer is to soften the copy, tweak the cadence, or add a privacy notice somewhere in the footer. None of that fixes the core issue. If the customer cannot reasonably understand why a message appeared, the system has already become too opaque.

The more invisible the logic, the more unsettling the result.

Privacy-first changes the narrative

This is where I think “privacy-first personalization” often gets oversimplified. It is not a nicer wrapper around the same machinery, and it requires different choices from the beginning.

A privacy-first model asks: what did the customer knowingly give us? In what context did they give it? How long do we really need it? Can we explain, in plain language, why this recommendation or message exists? Can the person change that experience without digging through a maze of settings or hunting down and directly emailing a data protection "officer"?

Those questions sound restrictive, but I think they produce better communication. They force teams to build systems that are more legible, more bounded, and less dependent on endless behavioral memory.

Smaller personalization is often better personalization

The future belongs to brands that know enough to stop before they cross a line. As legislation and regulation increase, so too does consumer awareness about how their data is used.

That means relying more on first-party context and less on sprawling data trails. It means responding to a moment rather than building an indefinite memory of every action. It means letting signals expire. It means accepting that not every possible data point should become part of a permanent profile.

What respectful personalization actually looks like

The standard should be pretty simple. Personalized communication should feel connected to a relationship the customer recognizes.

It should be understandable. It should stay close to the context in which the data was given. It should not rely on hidden leaps that make the brand feel like it is peering through the walls. It should also be reversible. If someone wants to reset, opt out, or correct the system, they should be able to do that quickly.

That makes personalization feel far more believable and meaningfully authentic.

And believable personalization is far more valuable than aggressive personalization that produces a momentary lift while quietly damaging trust.\

The better question

The industry has spent too long asking how far personalization could go. I think the better question is where it should stop. We live in a time of rampant, frantic, and potentially harmful data collection practices by companies that fewer and fewer people trust. Data stewardship has never been more important than it is now, and that's not going to change as time goes by, especially in an era dominated by AI innovation and the implications it has on privacy.

People see this. They see personalization and they find it uncanny. 

That is the real test of maturity. Not whether a company can build a system that knows more, but whether it can build one disciplined enough to know less.

Because the best personalization must feel like a brand paying attention within the boundaries of a relationship that still makes sense to the person on the other side.