Everything I’ve learned teaching SEO to university students in the age of AI

About five years ago I hired a young intern to help with what I’ll call the “grunt work” of much of my job then. Things like writing email campaigns, generating images for social media posts, etc. On his first day he sat down, and I said, “This company that makes fishing tackle needs an email campaign for Father’s Day. Here are the links to their site, here’s a shared folder with images, and here’s where you can draft the campaign.”

We nodded, he seemed to grok it, and then I turned back to whatever I was doing. A couple of hours passed and before I stood up ready for lunch I asked, “So what have you got?” 

He stared at his laptop blankly having accomplished precisely nothing. All he could say was, “I don’t know what to say.” 

In this moment I realized he had no concept of how to craft a message. Even one as simple as, “It’s Father’s Day, here’s 20% off,” was too much. 

“This isn’t good,” I thought. This set me on a path to thinking about why he had this particular challenge. He could manipulate his computer and poke around in Photoshop and seemed to have some understanding of basic website debugging. But nothing prepared him to write so much as an email subject line.

In conversations with local schools I started a process that wound through COVID and hiring freezes and the usual academic rigamarole and, this semester, I finally got in a classroom to teach a class on writing for the web.

Thing was, a new challenge arose in the last year: generative AI from ChatGPT and other large language models. It’s not really new this year, but it was new enough that it all collided with my initial class planning. The result has been an interesting semester that has forced me to reckon with what students do, what students know, what they should know, what I do, and how I do it professionally outside the classroom.

Confessions of an SEO man

For virtually my entire Internet-making life if you wanted to learn something, like HTML, Javascript, Flash ActionScript, how to structure a bunch of <div> tags, or configure a server, you could Google your way to it and there was no end to helpful, useful websites and places that would serve you examples. CSS-Tricks.com comes to mind as one genuinely helpful locale. 

But there was a hole: no one talked about how you do Search Engine Optimization (SEO). Everyone knew there was some kind of strategy because as long as Google has existed people have measured and finagled their way to the top. So much so even the most tech-illiterate clients would ask, “How do we get to the top of Google?” I’ve been asked this question for 20 years.

If you Googled, “How to make website rank in Google”, virtually nothing came up. Thumb through any digital subreddit and posts abound for “How do I learn SEO?” and you’ll find this is still the case today, though there are sites and services dedicated to SEO that purport to tell you how SEO works. But they’re frustratingly vague and don’t actually tell you how to do SEO.

The advice from Google that rattles around in most circles and has for years is some riff on:

  • “Write quality, authoritative content.”
  • “Provide helpful answers.”
  • “Make the best webpage on the Internet for your topic.”

This advice is not wrong. It’s even pretty good advice, and I’ve used Andy Crestodina’s “Make the best webpage on the Internet for your topic” as my Northstar for years. I’ve even repeated it in class this semester.

But … what does that even mean?

Because that sounds fine until you apply it to what can best be described as really banal, everyday stuff. Plumbing, house painters, wedding venue rentals, reviews about socks with just the right kind of fabric or stitching, marriage counseling, ideas for fundraising events, and all sorts of just…stuff. How do you make a “quality, authoritative” page that’s the “best webpage on the Internet” about a local florist?

The answer emerged with what became known as “content marketing.”

Anyone who lived through Friendster and LiveJournal and MySpace could see that social media platforms would come and go rather quickly, and they weren’t (and still aren’t) useful places to grow an audience. Not one that would support you for a long time anyway. Google was, and for now remains, the only place for sustainable, credible traffic flow.

So you start thinking up all kinds of stuff to go with the stuff about florists or plumbers. Here’s the trendiest flower for weddings in 2023. Here’s how to calculate how many grains your house needs for a water softener. And when you get here, here’s what grains are. You just have to start making stuff up with real answers, even if those answers are a simple one or two-line answer that you expound into 1,500 words.

The way you do SEO is simple: cobble together everyone else’s stuff and do it a little bit better

My crummy superpower in 2024 is I can take a topic and spin off a zillion questions about it. I am blessed to be just inquisitive enough that I can muse aloud on things that deal with boring stuff for everyday businesses and generate about 2 ideas a month every month…forever…and then write about them.

My SEO process works like this:

  1. Learn everything I can about the client, their business, and their work. Find out what makes them better and worse than competitors, then play to those strengths.
  2. Start picking some topics and ideas worth asking.
  3. Research the search volume of those topics in an obvious, Google-y way. Use a tool like Ahrefs to search the keywords, like, “trendiest flowers for weddings 2023” and see if it’s got a lot of traffic that’s both useful but not so hard it’s dominated by big players, like an HGTV.
  4. Start looking at all the other websites that rank for it and cobble together their headings, word count, use of images, and other media like videos. Prepare to outdo them either with length or volume.
  5. Identify a title with the keywords in it. It’s not, “Paint smell can last up to 48 hours” for a title, it’s, “Here’s how long paint smell lasts and 3 things that matter a lot.”
  6. Add 2-3 other sub-headings that relate to paint smell or flowers or whatever the topic is.
  7. Fill in text in between and add some stock images since no one ever has any good photos anyway.

That’s it. That’s the whole process. And I’ve gotten very good at it. Virtually every major site I work on today has the kind of hockey stick graph of traffic and site growth that everyone wants and imagines their site to have.

But I’m the SEO ghoul that Nick Heer writes about. This whole process is why recipe sites are a mile long, because there’s only so much you can say about an oatmeal cookie recipe. And it’s why the Internet has become festooned with “SEO content.”

Teaching this process to students reveals a potential future nightmare

I am old enough to remember a time when the web was dominated by websites like this one that you are reading right now. Bespoke little hubs of random posts and photos from their authors that reveal parts of their personality or a collection of interesting things. 

The university students of today have only fleeting memories of this. They were born after 9/11. Their nostalgia centers around the web of about 2007-2010, just as the iPhone was killing Flash and Ethan Marcotte coined the notion of Responsive Web Design.

Today I have students who use ChatGPT for form letters to me about missing class. I’ve received two this semester that were so plainly written by ChatGPT in the most stiff, formal tones it’s downright jarring. All to say, “I have a funeral to go to” or “I’m sick and can’t make it today.”

Here I thought the kid who couldn’t write a Father’s Day email promo was a problem. Now I have students turning in downright alien “content” they’re not even reading because it was just too easy to ignore ChatGPT.

I don’t blame them. Or ChatGPT. But if we think the web is lousy now. Just wait until people who don’t even remember the good times before the Internet infected with content start producing stuff en masse.

This is all primarily at Google’s feet, since they monopolized the web in search and when you start having benchmarks that generate money, people optimize for whatever the machine wants to make more money come out.

There is some good news in all of this.

  • Students tend to think, “No one reads,” but give them something compelling to read and they realize “No one reads boring things.
  • Young people recognize much of this is not good for them or anyone else. In fact, they seem downright terrified of AI.
  • Given a smidgeon of direction that, “You have to be better than AI because the AI is currently better than you,” it does inspire them to try harder. Or at least chew on that problem, which is a start.

The Internet may get worse before it gets better, but it will likely boomerang back around

Mark Twain’s old chestnut, “History doesn’t repeat itself, but it does rhyme.” comes to mind a lot. Students in both of my classes (I also teach a Photoshop class) have been asking me what, if anything, I think might happen or come of AI and how it will change industries. How terrifying it must be to be a young person right now facing the proposition that everything you’ve been studying will be automated away. Like you trained as an apprentice in buggy making at the same time Henry Ford rolled a Model-T down a dusty road in Michigan. Except now you see and know it’s happening at the same time.

I suspect the Internet, like businesses, will maintain lots of big top-tier brands that consume a lot of oxygen. Small businesses are going to have a harder go of it, but they always have, and the web is likely to boomerang back to a time when people checked certain sites a time or two a day. 

But we’re in it for the short term. It’s obvious to me that young people recognize they have problems relying too much on ChatGPT, TikTok, and other social media too much. But in the same breath will say, “I’m not going to change though. I like it.” Like an alcoholic who just likes being drunk. 

Developing personal editorial guidelines

Teaching this class has put me in the position of realizing there is no innovation in what goes into the web. Everyone is reverse engineering what already works and just trying to poke at the edges only far enough to nudge up in the same direction. Same goes for whether you’re writing webpages, producing YouTube videos, posting on Instagram, or whoring yourself out for karma on Reddit. 

I have numerous clients who pay me money precisely for my ability in getting their sites to rank for whatever they want them to rank for given time and effort. I get approached about twice a month from someone to do the same for them. But when I tell them what it takes some balk. “Oh, we don’t want all that on our website.”  Or, “We want a clean, minimal site,” (which is code for, “We don’t want anything on our website.” Then what’s the point?). I have no good answers for this — Google sets the way, and that’s the game we play. Yeah, it sucks, but people have to eat. What’s the point in making your little iOS app or web game or writing your poetry or news stories if no one even knows you exist and will likely never know

I am under no illusion that this post you’re reading right now will be read by mere tens of people. It’s not optimized to do anything. I’m writing it for myself. I don’t care about social media much, so I don’t “engage” there because I don’t think anyone cares what I think. I write this for myself and with the fleeting hope that maybe, just maybe, someone will stumble into it that finds value. But it’s unlikely, because Google will not be able to identify this as much about anything in particular.

Anil Dash said recently that “AI can’t have new ideas.” That’s true, since it just remixes what’s already there. So, yeah, we all have to have new ideas until we die. Hopefully they’re constant, consistent, and good enough people don’t starve in the meantime.

I’ve tried setting editorial guidelines for myself in the past. So much so that I’ve toyed with including a synopsis at the top of the “SEO posts” so people can get in and out without Google balking at the short page text (nothing less than 800 words is going far in Google, ever. Just ask the mommy bloggers sharing cookie recipes.) This feels like I’m dancing around the edges.

I need something more robust that matches my personal mission, vision, and purpose, and doesn’t make the web worse. So far, my guiding ideals are:

  • Be honest in every sentence.
  • Instead of just asking, “Is this the best webpage on the Internet about this topic?” ask, “Is this a new webpage about a new way of thinking about this idea?”
  • Does this page serve the reader with everything they need?
  • Does this page avoid mediocrity? Is it enjoyable?
  • Does it contain a compelling story that enhances the reader with knowledge and information that improves their understanding of the world?

I do not know how to fully implement this on banal posts about everyday businesses. Sometimes the answer to, “How long does it take to paint a house” is just “about 36-48 hours.” My job, it seems, is to figure out how to make that 998 words longer and enjoyable to read. And how to teach that to a new generation of creators and media artists who are lured by ChatGPT. I’m open to ideas.


Want to know when stuff like this is published?
Sign up for my email list.

Photo of Justin Harter

About JUSTIN HARTER

Justin has been around the Internet long enough to remember when people started saying “content is king”.

He has worked for some of Indiana’s largest companies, state government, taught college-level courses, and about 1.1M people see his work every year.

You’ll probably see him around Indianapolis on a bicycle.

Leave a Comment