Looks like Google's Review Rich Result is borked. Reviews with x/10 rating are being presented with a rating out of 5 instead. caranddriver.com/mazda/cx-70 "reviewRating": { "@ type": "Rating", "bestRating": 10, "worstRating": 1, "ratingValue": 8.5 }

Google has really cooked this core update on the geotargeting level. When you see US, UK and Indian sites ranking in AU results you know it's been borked.

Possibly the most detailed summary I've come across of the DoJ vs Google trial. Well worth downloading/reading. Very well written by Shaun Anderson. Only thing I'd ask for an index with links to each section.

Honestly, pretty disgusting: pressgazette.co.uk/news/google-ai… This is just about as close to outright theft as possible, given that Discover is "queryless". Google clearly has no boundaries left in pursuit of usurping traffic. Any lingering beliefs they care about publishers now gone.

Wonderful piece of sleuthing Metehan Yesilyurt: metehan.ai/blog/perplexit… I do have a question: the post appears to present as fact what each of these properties represents. Are you speculating, or do you have more data which explains directly what each property is doing? Thanks!

This is really interesting: tryprofound.com/blog/ai-search… The fly in the ointment is that this compares citations pre and during a core update. Citations are prob part of what Google refers to as "grounding", ergo update possibly responsible for at least some of Google's "drift".

AIO grounding fail: "The Garmin Forerunner 970 does not exist." (clicked on "TopoActive maps" in first AIO)

I did a LOT of work with font optimisation over the years, and this is a great overview. Covers just about every aspect of font optimisation you'll likely ever run into. The only trick I'd add are tools to match font vs fallback so that FOUT doesn't => CLS.

This is what I thought was happening. Google may try to block this. Doubt they'll let it continue without reply, and would be a valid strategic response. Could be quite nasty for OAI and Perplexity if Google does manage to block. Possible knock on effects for SEO industry though.

&num=100 support dropped Av. position soars Impressions dive Clicks stable Ergo low rank impressions went away Theory: Many rank trackers/scrapers pull &num=100 (incl. SEPRapi who serve OAI). Support pulled for strategic reasons. seroundtable.com/google-search-…

This snippet doesn't appear anywhere on that page. facebook.com/groups/HomeAss… Looks AI generated, but no indication of that on SERP if so.

Hey The Guardian I've worked on image cropping at scale, and know how hard it can be to address all edge cases, but this one is pretty bad to show on your Int homepage: Headless Kerry James Marshall

Gary's fingerprints all over this: 'And when we say "last modification", we actually mean "last significant modification".'

Nice to see some more theoretical SEO articles. This is a good read IMO: hobo-web.co.uk/e-e-a-t-qualit…

Pure hallucination. Actually, 2 pure hallucinations: google.com/search?q=schem…

Hey The Guardian Guardian Tech your takeover ads are totally ruining my web experience. The entire page gets downscaled when the takeover loads (video below, third refresh re-triggered the takeover). 4k monitor @ 150% scaling. Chrome, Win11.

Some great stats and analysis in here. Also goodbye Top Stories Carousel, right Patrick Stox? So lucky to be able to analyse big data like this (and we're so lucky it's shared):