Pinned Tweet
I built a RAG pipeline for SEO use cases with @zilliz_universe and @DeepnoteHQ
foreveresque.substack.com/p/…
Sharing a detailed step-by-step guide on why and how to set up your own RAG (with 0 cost other than the LLM API calls) to do SEO smarter not harder
ForEVer retweeted
Made a Chrome extension to bring semantic SEO to everybody for FREE!
Excited to launch our FREE Chrome extension that uses Chrome's built in gemini model to do semantic semantic analysis of websites. Extract keywords, score content, find the most relevant "chunk" and more! serprecon.com/chrome
ForEVer retweeted
I use this daily as part of my AI SEO workflow and decided to share with the rest of the community.
Article: dejan.ai/blog/grounding-snip…
Tool: snippets.dejan.ai/
TL;DR: Shows you which parts of the content were used to ground Google's AI search systems.
ForEVer retweeted
Tool: compressor.dejan.ai
Use: Enter URL to get 5 levels of semantic content compression.
BEFORE: "I just feel like this page lacks some love"
NOW: "Needs more contentEffort"
contentEffort is a Core Metric: The Google leak reveals the contentEffort metric, where LLMs algorithmically measure the human labour in your content. Effort is now a ranking signal. hobo-web.co.uk/the-contentef…
28
SEO hot take:
Knowing how to Google things was a superpower and now with LLMs and AI overviews it has been highly devalued
Whichever LLM interface people are using it is raising the skill of the average Google user
And when you acquire a skill you tend to want to use it more!
2
3
50
ForEVer retweeted
Anyone know anyone at Reddit?
Or can I borrow your Reddit API keys for a day? 🙏Really want to run something for BrightonSEO on Wed. ♥️
Love to think that a student or two did this with my key lol #SoProud
ForEVer retweeted
I do not subscribe to the classic Gartner hype cycle of tribal SEO bullshit
Not hiking along to the Peak of Inflated Expectations. Not digging your Trough of Disillusionment
Just adopting early and learning about what will be useful long after we hit the Plateau of Productivity
But pshhh don't tell any of your clients 🤫
Replying to @KorayGubur
Also the bottom positions don't have impressions anymore because of the num=100 change, so the avg position raised for everyone that used tracking tools for their sites 🤭
1
26
In SEO you should know what an experiment is and how to set 1 up even if you can't control all of the variables
You should also know how to write a hypothesis (hint: research any of the big 4 consulting firms and their formulas for a quality hypothesis within a business setting)
1
22
How is the great SEO de-noising treating y'all?
I for one am happy with less noise, same signal, can't beat the ratio
18
If anyone is feeling SEO nostalgic here's a bit of weekend reading for you to bookmark: a 2007 back-and-forth in the comment section, between @mattcutts and @JohnMu on how G should deal with paid-for links
mattcutts.com/blog/how-to-re…
😍
1
37
ForEVer retweeted
SEO News: if you missed my post over the weekend, there’s been some massive fluctuations to data in GSC.
Interestingly, this change may actually point to a broader underlying issue in our data that was originally thought to be related to the AI overviews rollout.
Impressions going up but clicks not following. This has been the shared understanding of the SEO community based on recent studies, but we may be now discovering how flawed this methodology was in the first place.
Google has been pushing back on these studies recently and the disabling of &num=100 may have motives outside of preventing competition from scraping their search results.
The outcome? Well, it seems like desktop impressions in particular in GSC for many sites (due to desktop being the primary rank tracking surface) may have been heavily influenced from an influx in bots (due to new AI analytics entrants) scraping their search results.
We can see this in the graph shown, where desktop impressions have dropped a whopping 200K impressions overnight for this site, forcing average position to move in a direction that is more close to mobile.
So you’re telling me 200K daily impressions on desktop were from scraper tools!? That seems a bit ridiculous, but potentially not fully outlandish. There is likely more to this story.
The reason for this is that by default, most users would have the top 10 search results shown in Search. An impression is only recorded in GSC if the result is loaded on that page and the link is within visible, but it doesn’t need to be scrolled into view. When the top 100 results are loaded, this will completely throw this dataset out of whack, causing the average position to increase dramatically.
Many think that Google has made this change to prevent external services abusing their systems. While this may be true, I do wonder whether there are ulterior motives in putting “the great decoupling” theories to rest.
















