Todd Heft

Last November, I ruminated on how generative AI was changing SEO. Well, my friends, the destruction of the SEO industry marches forth at blinding speed and is now nearly complete (local SEO is the exception - you still have options for now).

Sure, you can say that SEO pros should stick to the tried and true methods of content marketing, citations, backlinks, blah, blah, blah, and you'll be mentioned in the generative AI results that MIGHT get you a click or a mention here and there. But even that is wishful thinking, and I would say, a huge crock that ignores the speed of reality. I, for one, will not lie across the railroad tracks of established SEO methodology as the AI locomotive is bearing down on me.

Generative AI search results were incredibly shitty a year ago - really atrocious at times, and sometimes laughable at how wrong they were. I was invited to Google's generative AI search results beta way back when, and I thumbs-downed more results than I can count (and submitted explanations like someone who cares). I wasn't just trolling in fear of my career falling apart. But any search that required even infinitesimal expertise was woefully lacking. Now, even complicated concepts are somewhat reliable and growing more so every month (but always verify citations, because AI hallucinates - hello to that lawyer who files a legal brief and receives judicial wrath from a clear-eyed judge who says, yea, those are not real, and why didn't you check?).

I would qualify that - SEO as we knew it is most definitely dying and gasping for its last breath through a cocktail straw. One can no longer count on content strategy to drive keyword ranks in organic search, much less generate significant traffic. Even if you do get lucky with a piece today, it won't last. As AI gobbles up and regurgitates your IP, there is less and less reason for anyone to scroll down the page for an answer. A big part of this, as I've said before, is human nature - most humans, being the lazy monkeys we are, won't scroll if the answer at the top of the page sounds like it might be correct (even if it isn't).

Many SEO practitioners, hanging on for dear life, will recommend backlinks to establish brand authority and citations. This too is a fool's errand, IMO. Most backlink campaigns are insanely expensive compared to the quality of the links you receive. Paying thousands of dollars for links in forums on unrelated-to-your-vertical sites in Malaysia benefits no one except the agency that sold you that bucket of chum. For the price, one would expect eight links from the New York Times. Plus, if you don't know, Google is wise to this scam and devalues those links anyway.

All of this plays into the argument, "Are website traffic stats a vanity metric?" I agree with Rand Fishkin on this.

In the past, we all relied religiously on what Google reported in GA, and to a lesser extent, GA4. Careers were made on this data. Yet when one looks at organic traffic stats today, it appears that a lot of this traffic is driven by bots. Visits of less than 20 seconds are highly suspect and should be thrown out, which sinks your organic stats further. After all, if the user didn't actually engage with the page, the visit doesn't matter (hello, Meta and your 2-second visits you count as engagement). Therefore, it's a vanity metric because it doesn't reveal anything about the actual impact of your marketing efforts.

A large part of the organic problem is a plethora of AI-spun dreck and content that's been spit out by entry-level agency copywriters with no real expertise in anything (no offense, but youth has its drawbacks in this area). This perpetuates a cycle of mediocre or worse content building on more mediocre or worse content, which Google's AI digests and regurgitates as "the answer", without incorporating content from, oh, you know, EXPERTS. But this is now the world we live in so we all have to suck it up (granted, AI answers are slowly becoming more authoritative).

How do SEOs justify their existence? Some have now changed their pitch to GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), or Search Everywhere Optimization. And it can be really convincing. They use a lot of fancy language to tell you that your content needs to be structured in a certain way so the salient points can be extracted by generative AI where it can (possibly) be seen in answers and (possibly) clicked on. Yet, organic clicks and CTR keep dropping precipitously (see the research from Bain & Company and Search Engine Land below).

For all of these reasons, content strategy and content marketing shouldn’t be included in the SEO bucket anymore, because content should not be written to influence organic traffic metrics. Competent, experienced copywriters understand how to structure content to engage an audience, convert a customer, or impart information. By association, generative AI will extract the salient points for answers, based on the specific wording of the query.

Content’s focus should strictly be to touch your audience or customers wherever they live to further the customer journey or reinforce the brand. Write it for email, press, social, blog posts, what have you. But stop trying to convince anyone that you can optimize it for organic search, much less "answer engines", because that’s probably a game we can no longer win.

Naturally, technical SEO is still essential - having a website that's optimized for speed and UX is critical. And this highly specialized expertise should probably be the only thing that's known as SEO going forward (because it's truly optimization).

Resources:

How Generative Engine Optimization (GEO) Rewrites the Rules of Search, Andreesen Horowitz.

Goodbye Clicks, Hello AI: Zero-Click Search Redefines Marketing, Bain & Company

New data: Google AI Overviews are hurting click-through rates, Search Engine Land