Digital Muscle SEO agency

29/05/2026

🚨 Google just quietly dropped some really interesting Search updates, and I think these are worth paying attention to.
Yesterday, Google published a post about new features aimed at surfacing original, high-quality content in AI Overviews and AI Mode — and honestly, there's a lot to unpack here.
Here's what's new:

1. Preferred Sources is now inside AI Search.
This is a big one. You've been able to set Preferred Sources for a while, but now those sources will be labeled and highlighted directly inside AI Overviews and AI Mode responses. Google says people are twice as likely to click through to a Preferred Source — so if you run a site, you should absolutely be encouraging your readers to add you. There's even a documentation page for publishers on how to do this.

2. A new article carousel for developing topics.
When you search about a topic that's still unfolding, Google will show a more prominent carousel of timely articles alongside the AI response. This feels like Google acknowledging that AI Overviews alone aren't always enough — people still want to read something. This could be good news for publishers who cover news and trending topics.

3. Perspectives carousel.
Similar to the above, but focused on forums, social media, and online discussions. Think Reddit-style insights surfaced more prominently. This isn't entirely new territory for Google, but the prominence being given to it is notable.

4. "Highly Cited" badges.
This one's interesting from a quality signals perspective. Google is expanding the "Highly Cited" label to more web results — marking articles that other stories frequently reference. They're also noting when an article cites a Highly Cited source. To me, this looks like Google leaning harder into citation patterns as a trust signal. Original reporting matters.

My take: This feels like Google genuinely trying to thread a needle — keep users in AI-powered search while still driving traffic to publishers and rewarding original content. Whether that actually plays out in practice for your site is another question, but the intent behind these features is encouraging.
If you haven't set up Preferred Sources yet, now is the time: google.com/preferences/source
Would love to hear if anyone's already seen these features rolling out in the wild. 👇
🔗 https://blog.google/products-and-platforms/products/search/original-high-quality-content-search/

New updates to help you spot and connect with high-quality, original content in Google Search.

07/05/2026

Why did the SEO specialist panic when he heard about Hantavirus?

Because it was the only thing that could make his keyword rankings drop faster than a rodent-infested website in Google’s algorithm!

01/05/2026

SEO is a bit like this.... all of the time...

09/04/2026

Something I've been experimenting with lately: uploading Google Search Console data directly into Claude to help identify optimization priorities. I wanted to share what I've found because it's been genuinely useful, and I think a lot of site owners could benefit from trying this.

I'll walk through exactly how I've been doing it.
Export your last 3 months of GSC performance data
Go to Search Console → Performance → set your date range to the last 90 days → export as CSV. Three months gives you enough signal without too much noise.
Upload it to Claude and ask some specific questions
I've found the most useful prompts to be fairly direct:

Which queries am I currently ranking for, and what's my average position for each?

Which queries have strong impressions but a low click-through rate?

Which queries am I ranking between positions 11 and 20 for?
Based on this, where do you see the most realistic opportunities for improvement?

The key is that you're working with your own data — not industry benchmarks or guesses about what might be working.

Use the output as a starting point, not a finished plan
Claude is quite good at surfacing patterns — particularly those "page 2" queries where rankings sit between 11 and 20. These are worth paying attention to because the content already has some traction with Google; it may just need refinement.

I recently did this for a client site and found a cluster of 14 keywords sitting in that 11–15 range. We looked at those pages carefully — reviewing the content quality, relevance, and whether the pages were genuinely serving the searcher's intent — and made targeted improvements. Traffic to those pages increased noticeably over the following weeks.

I want to be clear that I'm not suggesting AI replaces careful analysis or an understanding of what Google is actually looking for in terms of quality. But as a way to quickly surface what's in your data and ask useful questions of it, I've found this approach really practical — especially for smaller sites that don't have large tool budgets.

If you try this, I'd genuinely be curious what you find. The patterns that show up tend to be quite site-specific, which is part of what makes it useful.

05/04/2026

Important: Google Search Console Has Been Over-Reporting Impressions Since May 2025

If you notice a drop in impressions in your Google Search Console Performance report over the coming weeks, there's no need to be alarmed. This is not a ranking issue.

Google has confirmed a bug in Search Console that has been silently over-counting impressions since May 13, 2025. A logging error caused pages to appear as though they were surfacing in search results more frequently than they actually were. Clicks and all other metrics were unaffected.

Google is now rolling out a fix. As corrections are applied, you will likely see a noticeable decline in reported impressions. This rollout is expected to take several weeks to complete across all accounts.

What this means for your data
Any impression-based reporting or KPIs covering roughly the past 11 months should be treated with caution. If you've been tracking impression trends, setting targets, or presenting data to clients based on Search Console impressions during this period, it's worth revisiting those numbers with this context in mind.
Clicks, fortunately, remained accurate throughout. If you need a reliable signal from this period, click data is your best reference point.

What to do now

Note the May 13, 2025 start date when reviewing historical data
Flag this for any clients or stakeholders who may see the drop and be concerned
Avoid drawing conclusions from impression trends until the fix has fully rolled out

Source: Search Engine Land, April 3, 2026

30/03/2026

I've been thinking a lot about the stat that's been circulating lately — that Google's search share has dropped from 89% to 71%. A lot of people in the SEO world are framing this as Google losing. But I'm not sure that's quite the right way to look at it🤔

What's interesting to me is that total search volume worldwide has actually increased by 26% since ChatGPT launched. So people aren't searching less. They're just searching in more places — Google, ChatGPT, Perplexity, Reddit, YouTube. It really does seem like search has evolved from a destination into a behaviour, one that's now distributed across a wide range of surfaces.

According to data from Graphite.io (March 2026), ChatGPT alone now accounts for around 20% of search-related traffic worldwide. And what I find really worth sitting with is how that traffic works. Google returns a list of results. ChatGPT tends to synthesise an answer — often pointing to one source, maybe two. That's a meaningful difference for anyone thinking about visibility.

For a long time, the goal was straightforward: rank in the top ten. But I think what we're seeing now is a genuine shift in what it means to be "found." It's less about ranking and more about being the answer that gets cited, referenced, or recommended — whether that's in a traditional search result or an AI-generated response.

The sites I'm seeing do well in 2026 aren't treating SEO and GEO as separate strategies. They're thinking about search holistically and asking: wherever someone is looking for this, are we the best answer?

I think that reframe matters quite a bit right now.

25/03/2026

This is something I've been watching closely, and I think it's worth taking seriously.

Google has launched what they're calling Personal Intelligence in AI Mode. On the surface, it sounds like another incremental update. But when you think through what it actually means mechanically, it's a meaningful departure from how search has always worked.
Two people can now search the exact same query and receive different answers — answers shaped by their Gmail history, YouTube watch behavior, Google Calendar, and the broader footprint they've built across Google's ecosystem. Same query. Different result. Depending on who's asking.

I want to be careful not to catastrophize here, because we're still early in understanding how this plays out in practice. But the underlying shift is real, and it's worth thinking through carefully.
Eli Schwartz flagged this direction back in September — specifically noting that Personal Intelligence, not AI Overviews, was the more consequential thing to track. His prediction was that AI Mode becomes the default Google experience at Google I/O in May. We'll see if that holds. But if it does, the traffic impact we've already been measuring from AI Overviews — which for many publishers has meant a meaningful reduction in organic clicks — may look modest by comparison.

Here's why I think that framing makes sense: GEO and AEO, as they exist today, still operate on a shared assumption. Optimize your content well, get cited, reach users. The answer might vary a little, but it's fundamentally the same answer for everyone querying that topic.

Personal Intelligence changes that assumption at the root. If Google is inferring intent from personal data signals before a query is even completed, the concept of "ranking for a keyword" stops mapping cleanly onto reality. There isn't a single SERP to optimize for. There's a personalized answer constructed for each user, built from signals that aren't visible to us and aren't something we can directly influence through content.

What I keep coming back to is this: the brands that appear consistently across personalized results are likely the ones Google already has familiarity signals for — through branded search queries, repeated user engagement, and the kind of accumulated recognition that builds over time. That's not a new insight exactly, but it takes on new weight here.

Brand has always mattered for E-E-A-T reasons. But in a world of personalized answers, it starts functioning less like a soft reinforcer of quality and more like a prerequisite for visibility in the first place.
I'm genuinely still working through what the practical implications are. But I think it's worth starting to think about how you're building that kind of familiarity now, rather than waiting until the shift is fully underway.

19/03/2026

I came across an AI-generated SEO audit this week that gave me pause.

The audit contained 60 action items — a mix of technical fixes, content recommendations, and general housekeeping tasks. On the surface, it looked thorough. But when I worked through it carefully, I had real concerns.

Several of the content suggestions, if implemented, could actually signal low-quality or manipulative content patterns to Google — the kind of thing that can contribute to a site being caught up in a helpful content or spam update rather than rewarded by one.

That's a meaningful risk that I don't think was adequately flagged.

The technical items were largely low-impact. I've seen sites spend months working through lists like this and come out the other side with no measurable improvement in organic traffic or revenue. That's a real cost — both in time and in opportunity, because every hour spent on a low-value fix is an hour not spent on something that might actually move the needle.

My concern isn't with using AI as part of an audit workflow — I think there's genuine value there when it's used thoughtfully. But AI output needs careful review from someone who understands why Google rewards certain sites and not others. Sending a raw AI audit to a client without that layer of expertise isn't a shortcut — it's a liability.

If you're working with an SEO consultant or agency, it's worth asking: what's the expected revenue or traffic impact of these recommendations, and how do you know? That question alone will tell you a lot.

17/03/2026

Organic search traffic to major publishers is down 42% since AI Overviews launched.

Breaking news is up 103%.

New data from Define Media Group tells an important story, but only half of it.

Yes, breaking news is thriving because Google has not (yet) let AI Overviews touch the Top Stories carousel. The speed and accuracy demands of live news are too much for LLMs right now.

This is great research if you are trying to understand the no-click landscape (in the comments)

20/07/2022

Remarketing 😅

The true story about how online advertising worksWant to promote your business effectively? Digital marketing agency Digital Muscle at your service h...