Google’s Preferred Sources Button Won’t Save Publisher Traffic

On Thursday, Google handed publishers something they’ve been begging for since AI search started eating their traffic: a button. It’s not a check for the clicks that vanished, and it’s not a way to opt out of AI Overviews. It’s an interactive “Preferred Sources” button that lets a reader tap it while on a publisher’s own website and say, this site is one I want to see more of across Search, Discover, and Google News.

I saw the news roll in through several outlets at once — TechCrunch, 9to5Google, Engadget all carried it within the same hour. As someone who runs a site that depends on search traffic, my first reaction wasn’t excitement. It was a slow, sinking realization that this is a band-aid Google designed to solve a wound Google itself opened.

What Google’s Preferred Sources Button Actually Does

Let’s set aside the reaction for a second and lay out what’s real. Google is now letting publishers embed an interactive button on their sites — as TechCrunch reported — that marks them as a “favorite source.” When a reader clicks it, that publication gets highlighted more often in Google Search results, in the Discover feed, and across Google News. The company tells us people are twice as likely to click through to a preferred source when it’s available — a stat it has cited in earlier rollouts.

This button is an expansion of the Preferred Sources feature Google rolled out in May to its AI experiences — AI Mode and AI Overviews — after it had already lived in the Top Stories carousel for a while. By the time of that May launch, Google said readers had already selected more than 345,000 unique sources through the same mechanism. That number tells you both how popular the idea is and how far it still has to go.

Alongside the button, Google is letting readers tweak their Discover feed in plain language — tapping a three-dot menu and typing, in their own words, what they want to see more or less of, per Google’s own developer documentation. Android users get to customize the audio daily briefings inside Google News too. It’s a coordinated push to hand consumers more control over their own feeds.

Why This Feels Like a Band-Aid

Here’s the tension that keeps nagging at me. Every AI-driven traffic story on this site has the same shape: AI search answers the question directly, the user gets the answer, and the publisher gets nothing. No click, no ad view, no subscriber. That has been the complaint from newsrooms and niche bloggers alike for the better part of two years now.

This button does not change any of that. It does not make Google send more clicks. It does not put the publisher back in the answer box. Instead, it asks the reader to do the work that Google’s algorithm used to do — actively remember, actively choose, actively keep a mental list of sources they trust. It shifts the load from Google’s ranking systems onto the individual human.

That’s not nothing, but it’s also not a fix. A reader who already knows and trusts Bleuken doesn’t need a button to find me again; they’d bookmark me or type my name. The reader who is being served an AI answer on my topic never gets to the point of clicking my preferred-source button, because they never land on my page to begin with. The button only reaches the audience that already came to you.

It’s a loyalty card, not a growth tool. And loyalty cards are useful — but they don’t bring new customers through the door, and they certainly don’t recover the foot traffic that the mall’s new automatic doors redirected somewhere else.

The Trust Problem Nobody Wants to Name

There’s a deeper issue hiding under the surface, and it’s about trust — both in the sources and in Google’s own judgment. We’re now nearly four years past ChatGPT’s launch, and the studies keep landing: roughly a third of newer web pages now show signs of AI authorship. Search results and AI answers are increasingly built on top of a web that is partly written by machines, and the trust problem runs both ways — fake think tanks are now fooling AI chatbots, and readers are getting harder to convince, not easier.

That’s the uncomfortable backdrop for a “preferred sources” button. On the surface it sounds like Google finally acknowledging that people want real, human-made journalism over a churned-out overview. But a preferred-source signal is only as valuable as the trust the reader already placed in a publication before they ever clicked the button. It doesn’t create new trust. It merely formalizes trust that already exists.

And here’s where I get genuinely uneasy as a technologist: handing readers more manual controls does not solve the underlying collapse of confidence in what the algorithm surfaces. A button nobody finds is just a marketing prop. The real test is whether Google still sends people to sources it hasn’t been asked to prefer. If the answer is “fewer than before,” then the button is a distraction from the actual decision Google made — to answer with machines and let discovery fend for itself.

Felix’s Take, Unfiltered

Let me be direct, because this hits close to home. I run Bleuken. I write most of my own post-mortems. I’ve watched AI overviews swallow search real estate that used to belong to independent sites like this one, and I’ve had to think hard about what a one-person publisher can do that a trillion-dollar model can’t.

The answer, I’ve concluded, is the same thing that always worked: write things that only a human who lives here could write, earn the reader’s trust once, and make that connection deep enough that they come back. Not because a button told them to, but because the content rewarded them the last ten times. My SEO work now starts with “what would make the reader bookmark this?” rather than “what would rank?” That shift has been the single most useful thing I’ve done since AI search rewired the ecosystem.

So I don’t think the Preferred Sources button is worthless. I think it’s a meaningful instrument for audiences that already trust you — which is to say, it’s a retention tool hiding in a growth-shaped box. If you’re a publisher that masters it, you bake a bit more resilience into your relationship with the readers you already have. If you’re hoping it replaces the traffic AI search took, you will be disappointed.

What Publishers Should Actually Do

If you run a site or a small publication, here’s my honest playbook for this — and it lines up with the strategy I’ve been testing on my own corner of the web.

• Add the button, sure. It’s low-cost and it strengthens the bond with readers who already value you. Think of it as a visible handshake, not a revenue engine.
• But don’t orbit your strategy around it. Build email lists, build community, build a reason for people to come to you directly. Own the channel, because you don’t own the algorithm.
• Double down on what machines can’t fake. Data points, local context, lived experience, an actual opinion. That’s the moat.
• Keep an eye on the trust math. In a web where much of the newer content is machine-made — and where entire archives are being pulled apart to feed the models — the human-made stuff becomes more scarce and more valuable. Be the scarce one.

Google’s button is a small move on a very large board. It doesn’t change the endgame — which is that search is no longer the reliable faucet of traffic it was in 2019. But it does confirm something I’ve believed for a while: even the company that built the sorting machine now admits the human touch matters enough to be worth asking for.

The industry is wrestling with the same trust question at a bigger scale — OpenAI itself hit the brakes on AI this week, and the fight over voluntary pacing shows how thin the safety consensus still is.

The question that keeps me up is simpler. Will the reader be willing to tell Google who to trust, or has the constant churn of AI slop already taught them not to trust what any feed puts in front of them? Because if the button works, it’ll be because readers still believe in something worth preferring. And if it doesn’t, we’ll all know exactly why, whether Google admits it or not.