Scrolled past this and got *really* freaked out for a second.
09.04.2025 01:50 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0@blairmacgregor.bsky.social
Independent search analyst. I help people navigate the SEO landscape in a post-AI world. 15+ year marketer. Long-suffering #isles fan. https://www.trusthetics.com
Scrolled past this and got *really* freaked out for a second.
09.04.2025 01:50 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0๐ Fun flashback. For those who followed me here who might not know, I owned myspacesurveyfun.com, which I started early that year (2006) and took it from 0 to 15,000 monthly visits in roughly 3 months.
21 on this list is probably "myspace surveys." If not, I've gotta think it was top 100 at least!
Google has declared, in the court of Google, that they are not violating Google's Site Reputation Abuse policies. Furthermore, Google is not trying to leverage Google's ranking signals to abuse its Google search rankings.
25.02.2025 14:43 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0Yeah. And liable to get worse. Costs will inevitably be passed on to consumers from OpenAI, Anthropic, and the rest. 'Plus' certainly won't be $20 a month forever. (Or the limitations put on it will render it useless.)
15.02.2025 17:54 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0In other words, time/cost to build has come way down but it's not zero. Which means there's always going to be room for people to build and charge for tools.
Moats also matter. Ahrefs/SEMRush are likely not at risk any time soon just because of the sheer data they've compiled & continue to compile.
There's also how much someone wants to spend maintaining a tool like what I built. Because it's a scraper, it's liable to need repair at some point and I've got to still troubleshoot that process with GPT or whichever LLM I've chosen to work with.
15.02.2025 17:27 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0Is this the end of SaaS as we know it? My (amateur) opinion is I don't think so; mostly because the limiting factor is still knowing and understanding the "specs" of what you want built and why.
15.02.2025 17:27 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0I just built two different SEO scraper tools with GPT, both in Python, the second using Selenium as a headless browser to scrape the data I need.
I can't code my *own* way out of a paper bag, for the most part.
Crazy times.
(Of course Google won't let anyone confirm or deny that assertion because no AIO data is being made available to site owners through GSC)
04.02.2025 18:08 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0Of course what these elements do is keep people on the SERP. The sources *are* cited but there's so much accompanying information, I'd imagine not a lot of people are clicking through to the publishers.
04.02.2025 18:08 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0Notice the flyout menus under each of these different points of comparison with long lists of specs. In this case, I'm not sure they add much additional value upon clicking. The fact that it's so spec-heavy makes me think of landing on an over-optimized product page.
04.02.2025 18:08 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0Another example of an AI Overview Google's experimenting with functionally taking up the whole SERP for a highly transactional "vs" keyword comparing two kids' bikes models.
04.02.2025 18:08 โ ๐ 2 ๐ 0 ๐ฌ 1 ๐ 0From a company perspective: after-action reports or debriefs? Feels like that's the closest equivalent in terms of documenting what went right, wrong etc. for next time.
31.01.2025 16:41 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0Google wants (or says they want) to reward writers & publishers with domain expertise who know their subject inside out. E-E-A-T!
At the same time, journalism itself is getting squeezed; the ones that remain are spreading themselves thinner, working multiple beats.
Seems bad!
More often though, in my experience, the 47 bosses, bosses' bosses and bosses' bosses' bosses needed to sign off on *one thing* lead to a stagnation within the system and all of the people involved (rightly or wrongly) giving up out of fatigue.
28.01.2025 17:19 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Everyone knows about the graveyard of completed-but-not-executed-on audits at the enterprise level, and sure, in some cases, the quality of the recommendations is sometimes just bad/clearly spit out of a tool......
28.01.2025 17:19 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0Plenty of reasons why legacy brands & the agencies that service them lose with SEO but speed is a big one. Take SERP features. Many of them move so fast at this point that if your average time-to-execute is 3-6 months, the SERP feature could be gone by the time you get approval.
28.01.2025 17:19 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0Google rep: Now that it's been a little over 18 months since we buried UA forever, how do you feel about GA4?
Me:
Oh *this* could be fun. ๐
28.01.2025 14:19 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0The weekly back and forth I have with myself between "any forecast has to quantify all possible scenarios with unassailable logic and point-by-point explanations of all formulas and factors ๐ ๐ซต" and "forecasting is all just guestimating, just do your best it's cool man โฎ๏ธ"
27.01.2025 16:48 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0I haven't tried DeepSeek yet (hopefully, I'll get a chance to play around with it today) but as of this second, I agree with this take: LLMs still aren't equipped to deal with tasks and workflows that require exact precision. But they're great for everything else. (And "everything else" is *a lot*!)
27.01.2025 14:49 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Not going to say it's *all* vanity traffic they lost/much ado about nothing (since eyeballing their CRM terms, it does seem like they've had some losses there).
But I'm also not going to make this into something it's not just to spin a narrative and send everyone into a panic.
hUbSpOt's dYiNg!!! SeO iS dYiNg!!!!11
Oh wait. Actually, it looks like the main domain's traffic stayed relatively steady at the same time a bunch of ToFU content on the blog subdomain (that probably contributed zilch to pipeline/revenue) tanked. ๐ค
If you saw even the slightest lift in CTA clicks, sales, leads, MQLs, SQLs or whatever other conversion metrics are important to your business from your already existing traffic, wouldn't that be worth it at the end of the day?
24.01.2025 14:43 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Even if you think E-E-A-T is BS and maintain these things don't do anything to directly lift SEO performance, think of it through the lens of CRO.
Say you're right. Say you did all this stuff, and it didn't move the needle in any way traffic-wise. You just maintained what you were already getting.
Off-site, If you mark up each author bio page with Author schema, include sameAs properties that include other profile pages and other publications they've written in to show that yes, they know what they're talking about and they're not some made-up, AI-generated person.
24.01.2025 14:43 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0On-site, flesh out your author bios. Real photos. Mention all their credentials; hell, show the actual degrees/certificates attained. Mention all of the places they've been written or had worked published.
24.01.2025 14:43 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0As LLMs gain more prominence and people continue to get fooled by AI slop, they will naturally start to distrust all or most written web content if it's unclear that the stated author is qualified to write about it.
24.01.2025 14:43 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0I know everyone's down on schema in general; and maybe rightfully so. But this is why, if you're playing in a space that requires high degrees of trust (YMYL or adjacent), it still pays to invest in anything that can help legitimize your authors in the eyes of search engines and readers.
24.01.2025 14:43 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0