The August spam update: what actually happened to food blogs

Google ran this update August 18 to 20, 2026. It finished in about two and a half days, which is fast. We watched it land across every food blog we measure. Here’s what it did.

PreliminaryPublished Aug 26, 2026 · data through Aug 23
Share

You’ve probably seen the scary posts already. Every Google update comes with a wave of “is your traffic about to crash?” content. We don’t have to guess. Phynite measures the same group of food blogs every day, so when an update rolls out, we can see exactly what it did across the whole group. Not anecdotes from Facebook groups. Real numbers.

The short version: this update was good for food blogs. Most blogs saw no change at all, and the blogs that did move mostly moved up.

Everything below is anonymous. We only ever show percentages, never anyone’s actual traffic.

First, what a spam update is (and isn’t)

A spam update isn’t Google rethinking how it ranks recipes. It’s a cleanup. Google goes after sites that break its rules: copycat sites that scrape other people’s recipes, junk sites pumping out hundreds of machine-written posts a day, and sites gaming links. If you write your own recipes on your own blog, this update wasn’t looking for you.

One worry keeps coming up: roundup posts. Linking to other bloggers in a roundup you wrote yourself isn’t link spam, and this update wasn’t aimed at it. Link spam means schemes: buying links, swapping links at scale, junk sites that exist only to pass links around. A genuinely helpful, well-written roundup is the kind of page Google wants more of.

So the real question isn’t “did I get hit?” It’s “did I pick up ground when the junk got cleared out?” That’s what we measured.

Most blogs: no change at all.

About 6 in 10 food blogs came through the update looking like any ordinary week. Their traffic wiggled the way traffic always wiggles, and nothing more. If your numbers looked normal this week, that’s the typical experience, and it means exactly what it looks like: nothing happened to your blog.

A real bump31%11% ordinary weekSteady (normal range)58%48% ordinary weekA real dip11%40% ordinary week
How blogs came through the update week, compared with how many move in an ordinary week. “A real bump” or “a real dip” means the change was bigger than that blog’s normal ups and downs.

The blogs that moved mostly moved up

3 : 1About 3 in 10 blogs got a bump too big to be normal wiggle. About 1 in 10 dipped. That’s three gainers for every decliner. And the lift grew after Google finished: the two strongest days in our data came after the update was done rolling out. That’s what it looks like when junk keeps draining out of the search results for a few days after the switch flips.

0%-0.8%+1.9%+1.9%+0.1%+3.5%+5.3%08-1108-1308-1508-1708-1808-1908-2008-2108-2208-23an ordinary weekthe update
Search traffic across all blogs, day by day, compared with each blog’s own normal. The grey line is an ordinary week earlier in August, shown for comparison. The navy line is the update and the days after it.

Recipes on page two gained the most

Google Search Console tells us where recipes rank. Here’s the pattern:

  • Recipes already in the top 3 stayed put.
  • Recipes sitting at spots 11 to 20 (roughly page two) moved up a little over one spot on average.
  • Recipes buried deeper than spot 20 moved up almost two spots.

That pattern is the fingerprint of a cleanup. The junk was never sitting at the top of the results. It was clogging up page two and beyond. Clear it out, and the real recipes behind it slide forward. The deeper the page sat, the more junk stood in front of it, and the more room it gained.

← moved up (gained)moved down →Top 3+0.214–10-0.3211–20-1.3121+-1.91grey = an ordinary week
How far recipes moved, grouped by where they started. Bars pointing left mean “moved up.” The grey bars show the same measurement in an ordinary week: almost nothing.

We also checked whether this is just fall recipes waking up for the season. It isn’t. The typical blog here gets only 6% of its clicks from fall and holiday recipes, and the picture looks the same with the fall-heavy blogs set aside.

So what should you do?

Honestly: nothing. There’s nothing to fix and nothing to chase. This update rewarded blogs for simply being real, and the best response is to keep doing what you’re doing: publish your own recipes, work your plan, and let the cleanup work in your favor.

If you’re worried about your own numbers, your dashboard compares your blog with its own normal week. That’s the only comparison that matters, not a screenshot from a Facebook group.

One number is still filling in. Search Console reports a few days behind, so the newest days of the traffic lift haven’t reached its click counts yet. Through the update window itself, clicks held steady while rankings moved underneath. That’s the usual order: rankings first, clicks follow.

What’s next

Once the dust fully settles (about two weeks), we’ll publish the complete report: who gained, who lost, which kinds of recipes moved, and what happened to real traffic. Same blogs, same measuring stick, no anecdotes.

How we measured (for the curious)

  • Which blogs count: member food blogs averaging at least 50 search visits a day (or 50 Google clicks a day) before the update. Very small blogs are left out because percentages on tiny numbers mislead.
  • Every blog is compared with itself: the same weekday averaged over the four weeks before August 18. No blog is ever compared with another blog.
  • “A real bump” or “a real dip” means the move was bigger than that blog’s own normal day-to-day variation (outside two standard deviations).
  • Sanity check: we ran the identical measurement on the week of August 11, before the update. That’s the “ordinary week” shown in every chart.
  • Rankings come from Search Console, weighted by impressions. A page needs at least 100 impressions before the update to count, and a blog needs at least 10 qualifying pages per group.
  • Seasonality: recipes carry season tags from Phynite’s content system. Blogs heavy on fall content (a small minority) are flagged in the analysis.
  • Privacy: blogs are anonymized before analysis. Only percentages and ratios appear anywhere. No one’s actual traffic numbers leave the raw data.
Share