In February 2011, Google rolled out an algorithm update called Panda. It affected roughly 12% of search queries in its first pass, and a lot of the websites that had spent years gaming SEO to climb the rankings watched their traffic crater overnight.
More importantly, it affected me. At the time I was part of a company’s SEO team, focused on writing a lot of that content that had now cratered, and was laid off.
Before Panda, ranking on Google was more of a checklist (e.g. keyword density, meta tag stuffing, internal linking ratios) where those factors served as proxies for how search engines thought people engaged with online content. While the Panda update wasn’t the first of its kind, it was certainly the biggest to shift the search algorithm toward reflecting actual human behavior.
Google trained its search engine to approximate actual human judgments, and site owners who had been writing for the algorithm suddenly had to write for people. Companies scrambled to start generating content people actually wanted; companies that were already prioritizing content that people actually wanted were fine.
Now, 15 years later, it’s happening all over again — this time with different acronyms.
First of all, call it whatever you want I guess (Answer Engine Optimization, Generative Engine Optimization, Large-Language Model Optimization, A.I. Search Engine Optimization) but from the user’s perspective, it’s nothing different than more traditional SEO. If I’m searching for something online, I go to a search engine; so in that sense, Google and Bing and Ask Jeeves (R.I.P.) aren’t all that different from ChatGPT and Perplexity and Claude. And we definitely don’t need another three-letter whatever. All of this is just optimizing the search experience, so let’s call it all SEO.
But second, it’s obvious that a lot of people talking about getting your content to show up in these A.I. engines weren’t around 15 years ago. Panda shifted results to better reflect human experiences instead of being a checklist of corporate tactics. So try whatever tips and tricks you want, the A.I. algorithms are going to gradually evolve to match what actual human people want; so if you just start now with generating content that actual human people want, you’re going to be fine.
To be clear, it is important to understand these A.I. engines. It’s useful to know that some initial research has found, “adding relevant statistics (Statistics Addition), incorporating credible quotes (Quotation Addition), and including citations from reliable sources (Cite Sources) in the website content, require minimal changes but significantly improve visibility in GE responses, enhancing both the credibility and richness of the content.” But if you had asked me 10 years ago how to make your content richer and more credible, I would have told you to add relevant statistics, incorporate credible quotes, and include citations from reliable sources.
It turns out those are just things that people like.
Panda was so disruptive because an entire industry had convinced itself that ranking signals were the goal, instead of being a proxy for the goal. The goal was, and is, people finding online content they actually want to engage with. The algorithm was just the layer in between, doing its best to figure out which content fit the bill. When the algorithm got better at its job, the proxies stopped being useful, and the content built entirely around the proxies stopped working.
The current generative engines are going to do the same thing. Right now we’re all still trying to figure out how people are going to use these A.I. tools, and the algorithms will continually evolve to reflect that. In the meantime, content that’s being optimized to be cited rather than to be useful is, eventually, going to register as the same kind of low-quality signal that Panda was built to suppress.
The signals these systems are eventually going to optimize for are the same signals every search system has optimized for since 2011: actual human behavior.
The companies that survived Panda were not the ones that figured out the new checklist. They were the ones that had been doing the work of building useful information, content, websites, and interfaces. And they were rewarded for it once the algorithm caught up.
There's no reason this time should be different. If you spend the next 18 months chasing GEO / AEO tactics that work in May 2026, the most likely outcome is that you'll do a lot of work, get a temporary bump, and then watch it evaporate the next time the underlying models update. If you spend the next 18 months making content people actually want to read on topics you actually understand, the generative engines will find you on their own schedule, and they will probably reward you for it.
This isn't a clever new insight. It's the same thing marketers have been told to do for decades. The fact that we keep needing to be reminded of it every time a new acronym or technology shows up is, frankly, the reason marketing has the reputation it does.
The Panda update was a correction for an industry that was focused on gimmicks instead of on quality. It’s happening all over again, but I’m sure this time around we’ll do it the right way.
Right?
If this resonates with you, consider subscribing. Each week I’ll share essays and commentary drawn from today’s marketing trends and my upcoming book Marketing Isn’t Special — all focused on improving how marketing is understood, measured, and valued.
If this doesn’t resonate with you, I would never begrudge someone wanting a less cluttered inbox.