Lily Ray 😏

Lily Ray 😏

@lilyray.nyc

SEO by day, DJ by night

3.9K followers213 following2.2K posts

One good way to show that you have no idea what you’re talking about is to compare SEO to GEO/AEO and put “E-E-A-T” as something that only matters in the GEO column.

The site I’m working on saw its average CTR in France drop from 1.63% to .53% and clicks drop by 80% after AI Overviews were launched there. 🫠

I’ve been seeing this a lot lately: hyphens ( - ) appear to be breaking AI Overview titles 😅 Shop the new Ray Introducing Ray (They’re both missing -Ban) 😅

At the #SEOIRL conference, @joyhawkins.bsky.social just showed an example of a local AI pack in the personal attorney category - one lawyer’s self-promoting listicle was used as a citation for every one of the firms listed, and it determined the ranking order too. Embarrassing for Google tbh.

Lily Ray 😏 reposted

WordPress SaaS marketers, take notes! The next massive Google update is around the corner. @lilyray.nyc just dropped a piece that every WordPress site owner, agency, marketers and solopreneur needs to read right now.

Lily Ray 😏 reposted

If there is one post you read today, make it this one from Lily. Excellent article covering the latest doc updates Google rolled out recently, how it signals a potential HUGE update coming tackling the gaming of AI Search, etc. So, will the update be separate, or be part of a broad core update?

Lily Ray 😏@lilyray.nyc

Just published my latest piece in Substack! If you’ve been paying attention, you’ve probably also connected the dots: All signs point to a major Google update on the horizon. open.substack.com/pub/lilyrayn...

Lily Ray 😏 reposted

Fantastic article

Replacing your CMS with AI-generated pages? Enjoy discovering why content management systems exist, one production bug at a time. You’ve just inherited a thousand publishing problems your CMS used to solve without you noticing. www.jonoalderson.com/conjecture/y...

I put together a new Substack explaining why I’m predicting that a massive Google update is imminent - perhaps similar to what we’ve seen with Panda, Penguin, the Helpful Content Update or March 2024 core update.

I believe they will be targeting low-quality, scaled AI content, tactics meant to manipulate AI search responses, paid/inauthentic mentions, and other deceptive approaches.

Google is now showing a lot more AI Overviews for branded queries (h/t Chris Long and DemandSphere) which means a few things (if it sticks):

View full thread

Very bizarre choice by Google this time and I’m not sure most Google users are going to like it… we have all been conditioned to search Google for brands as a way of finding their website for 25+ years…

Google is now showing a lot more AI Overviews for branded queries (h/t Chris Long and DemandSphere) which means a few things (if it sticks):

- multiple sites can appear in AIO citations for these branded queries - not just the brand itself - so other brands may start earning #1 rankings in GSC for branded terms that aren’t theirs (with abysmal CTR/traffic)

Pulling some updated numbers on the performance of self-promoting listicles in AI Overviews (~100 queries): In September (compared to June), Google's AI Overviews cite 38% fewer self-promotional 'best of' listicles, and they leave out the brands behind them 83% of the time, up from 69% in June.

Also, in 70% of cases, the self-promoting listicle is now helping competitors get recommended while the brand recommending itself is not part of the recommendation

This is interesting to see - a few sites hit hard by Google's updates 3 years ago (including the HCU!) are starting to gain back visibility in recent weeks & months. Yellow one is a former client from 2024-2025 👀

The sites getting hit by Google right now for scaling low quality AI content are actually producing many of the exact same page types and templates that companies scaled with freelance writers before 2023 😅 History repeats itself

Since LLMs construct multiple fan-out queries and use search engines to conduct these searches simultaneously, doesn’t that kind of imply that SEO is actually *more* important than it was before? 🤔 (Better coverage = better visibility)

For how much Google writes about “inauthentic mentions,” you’d think they’d be better at combating it. I’d say they’re currently doing worse (AIO/AI Mode responses) than the other LLMs on this front; the others have started providing disclaimers about spammy / promotional results.

I just checked a dozen unrelated terms in Google Trends (US, past 12 months) and every single one showed the same #1 region: Wyoming (The least populated state in the U.S.) Drill down, and it's all coming from Cheyenne - a town of 67K.

View full thread

Which would line up with something I've been wondering about for the past year: how much of the unusual spikes in search volume we've been seeing in Google Trends (the "SEO" surge over the past year, for example) is actually coming from real humans.

Try it yourself: check any term in Google Trends in the US for the past 12 months. Curious whether you get Wyoming too.

I just checked a dozen unrelated terms in Google Trends (US, past 12 months) and every single one showed the same #1 region: Wyoming (The least populated state in the U.S.) Drill down, and it's all coming from Cheyenne - a town of 67K.

Even if Google Trends normalizes by a region's total search volume, not by population... that wouldn't explain why Cheyenne, Wyoming leads in search demand for *every query* I checked. Unless Google Trends data is inaccurate?

Cheyenne, Wyoming also happens to be a data center hub (link in comments). So perhaps the explanation might be that a lot of non-human traffic is coming from that location?

Despite reading AI search and SEO content every day, I rarely read any new finding or discovery that results in actually doing something new or different, strategically.

I'm pretty confident these are ChatGPT fan-out queries in Google Search Console. You can use this regex filter to check your own sites: (?i)(site:.*official|official.*site:)

It checks for site: and "official" at the same time. It's not *all* ChatGPT queries, but it's the ones that contain both of those terms.

I just used this across a few sites and I'm seeing the same pattern of many impressions, 0 clicks, and a whole lot of queries that look exactly like ChatGPT fan-outs in GSC.

As search becomes increasingly cross-channel and multimodal, it's a good idea to consider updating your organic search reporting to include more data than just your own website's organic traffic.

Since Google is now giving GSC data for various social properties, and external tools like Substack also provide GSC stats, you can combine all the data into a holistic organic report that shows how visible your brand is across the ecosystem.

Especially given that Google shows social posts in the "What People Are Saying" carousel for many high-volume queries... you'd be surprised by the total impression count when you add up this data across platforms.

Hey @johnmu.com I saw you recently say that there’s nothing special you need to do for AI search, but I have a feeling Zeeshan might disagree with you. Is Google actually planning to do anything about “inauthentic mentions”? I’ve seen some changes but this stuff can still work well.

ChatGPT is doing fan-out searches using site: … but it doesn’t always get the domain right. And even when it does, a site: search for a brand doesn’t just pull up the root domain. It can pull up all the subdomains too. 😅

View full thread

Of course, this is one of those things that, if it works, will inevitably be spammed heavily. And then OpenAI will need to figure out a new solution besides running site: searches across the entire domain and all its subdomains.

Like… canonical URLs? Building a search engine? 😂 By the way, this is NOT a recommendation to go spamming things. (Who, me? Never!) But I do think some cool SEO experiments could come out of this. (Proceed at your own risk!)

ChatGPT is doing fan-out searches using site: … but it doesn’t always get the domain right. And even when it does, a site: search for a brand doesn’t just pull up the root domain. It can pull up all the subdomains too. 😅

View full thread

But there might be some unique SEO testing opportunities with this new change. For example, what happens if you create content you specifically want to appear in LLMs on a subdomain of the main site?

ChatGPT is doing fan-out searches using site: … but it doesn’t always get the domain right. And even when it does, a site: search for a brand doesn’t just pull up the root domain. It can pull up all the subdomains too. 😅

I’ve already worked on a client project where I found that ChatGPT was doing site: searches that pulled up a subdomain every time, instead of the brand’s root domain… until we discovered that the root domain was accidentally blocking OpenAI’s web search crawlers. 🫠

Between ChatGPT heavily using site: searches in fan out queries (which prevents lesser-known domains from being retrieved there), elevating .gov, .org and other “official” sites in fan-outs… (1/2)

and now citing Reddit and YouTube less frequently and official brand sites more… this all sounds to me like OpenAI is trying to reduce spam making it into ChatGPT’s responses.