What I Learned: SEO Is Not Something You Hack

Over the past couple of weeks, I've been learning more seriously about SEO. One of the books I read was How to Get to the Top of Google, and it changed how I think about the whole subject.

My main takeaway is surprisingly simple:

SEO isn't something you hack.

You can take an incredibly data-driven approach. You can research hundreds of keywords, analyse competitors, automate content production, obsess over rankings and build sophisticated systems around all of it.

And your results can still be poor.

Because none of those things answer the most important question:

Why should anyone come to your website in the first place?

Start with something worth finding

For a service-based business, I think this is where SEO really starts.

You need genuine reasons for people to visit your website, trust you and potentially recommend you to somebody else.

That might be your expertise. It might be an unusual insight you've developed through your work. It could be a genuinely useful resource, an interesting programme you've run, original research, a survey you've conducted, a tool you've built, or simply an exceptionally good explanation of a problem your customers regularly encounter.

One useful question is:

What would make a journalist want to contact you?

Not because getting into newspapers is necessarily the objective, but because it's a good test.

If you've done something interesting, discovered something surprising, collected useful data or developed genuine expertise on a subject, other people have a reason to reference you.

Now you have something to work with.

Then SEO becomes much simpler

Once you have something worth finding, the more conventional SEO work starts to make sense.

You optimise the website so that it's fast, usable, technically sound and easy for search engines to understand.

You do keyword and search-intent research to understand what people are actually looking for and the language they use when looking for it.

You use that research to decide what pages and content you should create, and how to structure them.

You publish useful content around the areas where you genuinely have something to contribute.

And you promote that work. You contact relevant websites, publications, organisations and people who might actually find it useful. When you've created something worth referencing, outreach for backlinks becomes a much more natural proposition.

There's plenty of complexity within each of those activities, of course. But at a strategic level, I'm increasingly convinced that most of the rest is noise.

This changed how I think about AI-generated content

Before learning more about SEO, I was particularly interested in how much of the process could be automated.

And technically, quite a lot of it can be.

AI can research topics, help identify keywords, produce drafts, restructure existing material and dramatically reduce the amount of work involved in producing content.

But there's a problem.

If the underlying thinking isn't particularly valuable, automating its production doesn't make it valuable.

Suppose I'm working with a therapist. I'm not a therapist.

I could build a system that generates hundreds of articles about anxiety, relationships, depression or other subjects associated with their practice. It might even produce perfectly competent articles.

But where is the therapist in any of that?

A much stronger approach would be to start with the therapist's actual expertise.

What have they learned from years of practice? What questions do people repeatedly ask them? What do people commonly misunderstand? What patterns have they noticed? Are there useful resources they wish existed? Is there a subject they understand particularly well? Could we conduct a survey and ask a genuinely interesting question?

Those are much harder questions to automate.

They also seem far more valuable.

Once we have the answers, AI becomes useful again. I could interview the therapist, for example, and use AI to help turn that conversation into articles, guides, FAQs and other material.

In other words, automate the production where it makes sense, but don't try to automate the expertise.

SEO is more collaborative than I thought

This realisation has also changed how I think about offering SEO as a service.

I originally imagined it as something relatively straightforward that could sit on top of a website I'd built for a client: optimise the site, research keywords, create content, work on backlinks, measure the results and repeat.

I don't think I'd sell it that way now.

Before doing all of that, I'd want to sit down with the client and work out what we actually have to work with.

What expertise exists inside the business?

What could we create?

What could we research?

What could we do that would be genuinely interesting or useful?

What might give somebody a reason to mention this business, link to it, recommend it or come back to its website?

That requires substantial input from the client. I can bring the technical knowledge, research, marketing thinking, website work, outreach and execution. But I can't manufacture years of domain expertise on their behalf.

The strategy has to combine both.

And only once we've figured that out does it make sense to build the SEO machine around it.

This is really just marketing

Perhaps the biggest thing I've learned is that none of this is particularly unique to SEO.

It's marketing.

Have something good.

Understand the people who might want it.

Give them useful reasons to pay attention to you.

Demonstrate that you know what you're talking about.

Do interesting things that other people have reasons to talk about.

Then make it as easy as possible for those people to discover you.

SEO adds a particular set of tools and constraints to that process. GEO may change some of the ways in which information gets discovered. Search engines and AI systems will continue changing too.

But I suspect the underlying principle is much more durable:

First, create something worth finding. Then optimise how it gets found.