Why is news personalized?

News personalization isn't an accident. It's a deliberate product decision driven by specific economic and technical incentives. This page lays out the structure of those incentives, then explains what's lost in the bargain.

The economic argument

Most "free" news products are funded by advertising. Advertising rates are determined by engagement metrics — clicks, dwell time, return visits. A feed that holds you in the app longer earns more per session.

Personalization is the most reliable way to extend session time. Show readers more of what they previously engaged with, and they engage longer. The model is trained on engagement; the optimization target is more engagement; the result is a feed shaped to the reader's prior taps.

The economic logic is so strong that almost every advertising-funded news product converges on a personalized feed. The exceptions — wire-service homepages, traditional newspapers' websites — usually have non-advertising revenue (subscription, public funding) so they can afford editorial-led ordering.

The technical argument

Personalization at scale requires:

• An identity for each user (account, device id, fingerprint)
• An event log of every interaction (click, scroll, dwell, share)
• A ranking model trained on the event log
• A serving infrastructure that scores candidate items per user per request

None of these are trivial. Building them requires both engineering investment and user-data collection. Once a company has built the infrastructure, the marginal cost of running personalization is low — so the equation tilts toward "run it, since we have it."

The reader-experience argument (the strongest claim)

Defenders of personalization argue that readers genuinely benefit. The argument: news is a firehose, readers can't process all of it, the algorithm helps surface what's relevant. This is true in narrow cases (the New York Times' 200 daily articles really are too much for most readers to scan).

The problem is that "relevant" gets defined as "engaging," and engagement is downstream of priors, not informedness. The algorithm shows you more of what you already want to see, which over time narrows your feed and your worldview.

What you give up

Three losses worth naming:

1. Shared reality

When every reader sees a different feed, the public conversation fragments. You and a colleague reading the same news app on the same morning see different headlines. You stop being able to assume shared awareness of basic events.

2. Editorial accountability

When the order is opaque and per-reader, the publisher loses a meaningful editorial voice. The front page used to be a statement; in personalized feeds, there's no front page. The newsroom's editorial decisions are filtered through a model the reader can't see and the editor can't shape.

3. Discovery outside priors

The algorithm reliably under-shows you stories outside your prior interests. If you've never clicked a foreign-policy story, you'll see fewer of them. If you've never engaged with a center-right opinion piece, you'll see fewer of them. The narrowing happens slowly enough that you don't notice; the cumulative effect is large.

The alternative

The non-personalized feed is honest about being a firehose. It shows you everything that landed, in publish-time order, and trusts you to do your own filtering with category chips, source filters, and the discipline of skipping headlines you don't care about. WeSearch is built on this trade.

It's a less convenient experience — you'll see headlines you'd have skipped. The argument is that you'll also see headlines that turn out to matter even though you wouldn't have predicted you'd care about them. Whether that's a worthwhile trade is a personal judgment; we made one.

The deeper structural problem

Personalization isn't just one product feature among many; it's the central feature of advertising-funded internet companies, and removing it would unwind the business model. Google, Meta, TikTok, and the news products downstream of them all depend on personalization-driven engagement to sell ads. A news app that genuinely de-personalized its feed would forfeit a meaningful portion of its revenue. That's why the major aggregators almost never offer a clean "off switch" — even when they offer one (Apple News' "Today" tab, Google News' "Headlines" tab), the off switch is harder to find than it should be, and the algorithmic version stays the default.

The deeper claim is that the personalization economy isn't going to fix itself. Readers who want non-personalized news have to seek out specific products that explicitly refuse personalization, because the default for everything else is personalization-by-design. Wesearch is one of those products. RSS readers are another. Hacker News is a third. The list is short because the business model only works for products that aren't trying to maximize advertising revenue.

What this looks like for the news industry

The personalization-driven news economy has had visible effects on how news is produced. Headlines optimize for click-through-rate. A/B-tested thumbnails replace photo-editor judgment. Story selection bends toward what the platform algorithm rewards. Wire-service originality (the same factual story everywhere) gets passed over for "takes" with stronger emotional valence. The drift compounds across years; the news that surfaces in 2026 is meaningfully different from the news that surfaced in 2010 in ways that aren't fully explained by changes in the world.

This isn't a moral indictment of any individual journalist or editor. It's a description of how an entire industry's incentive system has shifted under personalization pressure. Reversing the drift requires alternative funding models that don't depend on engagement-driven ad revenue.

Bottom line

• News is personalized because advertising revenue requires it.
• Personalization narrows what readers see, fragments shared reality, and bends the news industry's incentives.
• The fix isn't an off-switch in a personalized app; it's a different product whose business model doesn't require personalization.
• WeSearch, RSS readers, and a few other products are explicit alternatives. Most mass-market news products are not.

Common questions

Frequently asked