What No One Tells You About Personalization
Source: https://uxdesign.cc/what-no-one-tells-you-about-personalization-150c9064ae8e ↗
Polonski's essay argues the uncomfortable point behind "personalisation" as a product feature: most implementations reduce user agency rather than extend it, and many produce filter bubbles and perceived surveillance without delivering obvious value.
The piece is a useful counterweight to the ambient enthusiasm about personalisation that dominates product conversations at scale.
For product direction the essay is a diagnostic prompt — before any personalisation feature, ask whether the user's experience of the product becomes richer or narrower, more transparent or more opaque.
Short, concrete. Pair with Chayka for the consumer-side account of the same dynamics.
Central argument
Polonski argues that personalisation as typically implemented is not a user-empowerment feature but a constraint: algorithms narrow what users encounter rather than expanding it, creating filter bubbles that limit exposure to diverse content and perspectives. Simultaneously, users experience the inference machinery as surveillance — the product appears to know too much — without receiving proportionate value in return. The central thesis is that the gap between personalisation's promise (relevance, delight) and its reality (opacity, narrowing) is a design and product accountability failure, not merely a technical limitation.
Critique
The essay was written in 2018 and draws heavily on filter-bubble anxieties that were prominent at the time but have since been complicated by empirical research — notably work suggesting that algorithmic feeds can expose users to more ideological diversity than their organic social networks do. Polonski's argument risks treating the filter-bubble effect as settled when the evidence is more contested, which weakens the diagnostic force of the critique for product teams who want empirical grounding rather than a persuasive essay.
Why it matters for product
For a CPO, the essay reframes how personalisation features get scoped and measured: if your success metric is engagement or CTR on recommended content, you are structurally blind to the narrowing effect Polonski describes, which means the metric itself encodes the problem. The transparency question also has direct implications for product strategy — features that personalise without surfacing the rationale to users create trust debt that compounds over time, making it a risk to account for in roadmap prioritisation, not just a UX nicety.