Food blog traffic is falling to AI Overviews. Here’s what to do next

What has happened to recipe traffic

Google started showing AI Overviews to everyone in the US on 14 May 2024. A year later, on 20 May 2025, it opened AI Mode, a chat-style search tab, to all US users. Both answer the question on Google’s own page, above the links.

For most topics that costs some clicks. For recipes it can cost the whole visit, because a recipe is a list of ingredients and a list of steps, and a summary can show both.

The reporting since then is consistent in direction, even if the size varies a lot from site to site:

• Pew Research Center tracked the real browsing of 900 US adults in March 2025. When a Google search showed an AI summary, people clicked a normal result on 8% of visits. Without one, they clicked on 15%. They clicked a link inside the summary itself on just 1% of visits.
• Ahrefs compared 300,000 keywords and found that an AI Overview went with a 58% lower click-through rate for the page ranking first (December 2025 data, published February 2026). Its earlier study, a year before, put the drop at 34.5%. The effect gets smaller further down the page but doesn’t disappear.
• Bloomberg interviewed 22 independent food creators for a piece Fortune republished in November 2025. They described Google traffic falls of 30% to 80%. Carrie Forrest of Clean Eating Kitchen said she had lost 80% of her traffic and revenue in two years and gone from about ten staff to none. Eb Gargano of Easy Peasy Foodie said traffic to her turkey recipe was down 40% on the year before.
• Chartbeat data reported by Axios in March 2026 found small publishers had lost 60% of their search referral traffic over two years, and medium publishers 47%. Chartbeat mostly measures news sites, so treat it as a signal about the open web rather than a recipe-specific figure.

The other side of the argument

Google disputes that this is a general collapse. In August 2025 its head of Search, Liz Reid, wrote that total organic clicks from Google to websites had been “relatively stable” year on year, while acknowledging that traffic is shifting between sites and some will see less. Google didn’t publish the numbers behind that claim.

Some well-run food blogs are holding up better than the headlines suggest. Pinch of Yum published its own referral mix in April 2026: search went from 61% of its traffic in March 2024 to 57% in March 2026. That’s a share, not a visit count, so it doesn’t show total traffic, but it isn’t a collapse either. Over the same period its email share rose from 3.9% to 6.8% and Facebook from 0.6% to 6.4%, after it started posting there three times a day.

So the fair reading is this: AI answers reduce clicks on many recipe searches, the damage is very uneven, and sites with a recognisable name and other channels are less exposed.

Why AI answers hit recipe blogs harder

The answer fits in the summary

Someone searching “banana bread recipe” wants quantities, an oven temperature and a time. AI Mode can show all three. Food creators call the result a “Frankenstein recipe”: in the Bloomberg reporting, Adam Gallagher of Inspired Taste described Google combining his ingredients with another blog’s method, even on searches for his own site’s name. Gargano found an AI summary of her Christmas cake giving a baking time close to four hours.

Google has adjusted the product more than once under pressure. It ended its Recipe Quick View test in July 2025; Bloomberg reported that participating creators had been paid, but not enough to cover lost ad revenue. In March 2026 Robby Stein, Google’s VP of Product for Search, announced that tapping a dish in AI Mode would now show links to recipe sites. Those changes help, but they don’t restore the old deal where every answer meant a visit.

Ad income is paid per visit

Most full-time recipe blogs earn through ad networks such as Mediavine and Raptive, which pay on page RPM: revenue per thousand pageviews. Fewer visits means proportionally less money, with nothing left over from the reader who got the answer on Google.

Rates have been falling too. In a Reddit AMA in September 2025, Mediavine’s CEO Eric Hochberger said CPMs were down on the previous year across the industry, with reported falls of 10% to 20%. He also said that “relying solely on Google Search traffic, as we’ve done for years, is no longer viable.” That’s the ad network saying it, not a membership platform.

A worked example

These figures are illustrative. Use your own dashboard numbers.

	Before	After
Monthly pageviews	300,000	228,000
Share from Google	60%	47%
Page RPM	US$30	US$27
Monthly ad income	US$9,000	US$6,156

The “after” column assumes Google sends 40% fewer visits (180,000 down to 108,000, so total pageviews fall by 72,000) and RPM drops 10%. The combined result is a 32% fall in income, about US$34,000 a year, from two changes that each sound survivable on their own.

What to do about it

None of this means abandoning search. It still brings people who have never heard of you, which no other channel does as cheaply. The aim is to stop each search visit being the end of the relationship.

1. Treat every visit as a chance to get an email address

An email list is the one audience you own outright. Nobody can summarise it, rank it or throttle it. Pinch of Yum’s email share nearly doubled in two years, and it was one of the few channels that grew for them.

Practical steps:

• Put a specific offer on your highest-traffic recipes: a free three-recipe weeknight pack, a printable shopping list, a holiday baking timeline. “Join my newsletter” converts far worse than something useful.
• Send at least one email a week. Mix new recipes with a short personal note. It’s the personal part an AI summary can’t reproduce.
• Track list growth as a headline number next to pageviews.

2. Sell something

A reader who gets value from you for free will often pay for something more organised. The common first products for food creators are:

• A themed e-book or recipe pack. Low effort if the recipes already exist, and it tests whether your audience will pay at all.
• Meal plans. Weekly plans with a combined shopping list solve a real problem: deciding what to cook. A blog post can’t do that.
• A seasonal guide. Christmas, Ramadan, back-to-school lunches. Time-bound, easy to promote by email.

3. Build a membership

A membership turns your best readers into recurring income that doesn’t move with Google updates. It works for recipe creators when it gives members tools, not just extra recipes: servings that scale, a shopping list that combines ingredients across recipes, a weekly meal plan, a cook mode that keeps the screen on. Those are things a summary at the top of a search page can’t give them.

The maths is small-number maths. Using the example above:

• Email list of 20,000 people.
• 2% join a US$8 a month membership. That’s 400 members, an assumption you should test, not a benchmark.
• 400 × US$8 = US$3,200 a month before card fees and platform costs, or roughly US$38,000 a year.

That’s about the size of the ad income lost in the example, from a few hundred people. They also don’t disappear when an algorithm changes. You’ll lose some to cancellations each month, so the work is keeping the plan worth paying for.

For the detail, read how to start a recipe membership and how much to charge for a meal plan membership. You can also model your own numbers with the membership revenue calculator.

4. Make your name the search term

Searches for your name, or “your name + dish”, are the ones you’re most likely to keep. Gallagher’s experience shows AI Mode doesn’t always respect them, but a reader who already knows you is far more likely to click through, or to come straight to your site or app. Put your name and face on your recipes, sign your emails, and give your signature recipes names people remember.

5. Spread the risk across channels

Pinterest has been no safe harbour. Colleen Milne of The Food Blog told Bloomberg her monthly Pinterest views fell from about 1.3 million to 419,000. Facebook worked for Pinch of Yum once they committed to it. The lesson isn’t which platform wins this year. It’s that no single source should carry more than half your income, and the sources you own (email, members, buyers) are worth more per person than the ones you rent.

A sensible 90-day plan

1. Weeks 1 to 2. Pull 24 months of Search Console and ad network data. Find the 20 recipes that bring the most traffic and income.
2. Weeks 3 to 4. Add a specific email offer to those 20 pages. Start a weekly email if you don’t send one.
3. Weeks 5 to 8. Package your best recipes into a paid e-book or meal plan and sell it to your list. Count buyers, not page views.
4. Weeks 9 to 12. If people buy, plan a membership around what they bought, priced for your audience. Launch to the list first.

Where Voyage fits

Launch with Voyage is how Voyage, a product and growth agency in New Zealand, partners with established food creators to launch their own recipe membership: a branded website and member area with scaling recipes, shopping lists, meal plans and cook mode, under your name, with an optional app of your own. There’s no monthly fee and no setup fee; we take a share of sales, from 10%, and members pay into your own Stripe account. If your blog already has an audience and you want income that doesn’t depend on Google, see recipe memberships with Voyage.