Effectiveness of automatic SEO apps

We all love it when things are done for us right?

Automatic washing machines, automatic transmissions, automatic vacuum cleaners...

SEO apps, well that's tricky. As one Shopify merchant asked:

How effective are apps that claim to "automate" your store's SEO by optimizing tags, alt texts & links?

Let's put "effective" aside for now and look at how well Shopify apps can automate tags, alt text, and links.

These are all easy SEO tasks. Low-hanging fruit that any SEO worth their salt can do in their sleep. Also all of those are just code (text) in your HTML and software is great at working with text. Software can very easily add text to the HTML of your pages so there are specific HTML tags, alt text, and as many links as you want.

Isn't that the promise of LLM AIs, they can create unlimited text content for you to use?

Sounds great right?

If it weren't for that pesky "effective" word though.

Since effective means having the intended effect and since this is about SEO, that means we want all of that added text to improve SEO.

Unfortunately, there is no magic text to improve your SEO. Sure there are rules of thumb such as use heading tags. But they don't always apply and may apply differently in different contexts. Reducing them into terms for software to do is like reducing a doctor's role to just prescribing medicine ("which medicine?" "doesn't matter") or a surgeon's role to cutting people open.

Adding text automatically is easy.

Adding the right text is hard.

The effectiveness comes from knowing how to apply the knowledge. For doctors that's pairing which medicine to use for which aliment. For SEO that's knowing which technique to use to get a specific result. Applying the knowledge is more important than the technique.

For doctors there are only a limited number of aliments they'll see, say a million. They might interact oddly but there's still an upper limit. For HTML code and SEO, there are an infinite number of combinations of text and code. That means true automation to cover every single combination is literally impossible.

This is why those automation tools can never keep promises of perfect results. They only have a handful of techniques coded for that they can use.

For example adding alt text to an image means adding a text description of the image. Some tools will just use the filename. So you might get something like "Times most influential images 10" from the following image.

An AI tool might pull out "Man kissing a woman in a street"

A SEO working on the image would instead label it as "World War 2 sailor kissing a nurse in Times Square to celebrate the end of the war"

(Photo credit: Alfred Eisenstaedt)

If those tools do so poorly on images, imagine how they'd mess up when all of your content, HTML code, and links are up for changes.

If you want to use automation tools, including AI LLMs, you have to review everything that comes out from them. You may get a lot of crap results but that's the price you pay with automation tools.

The biggest issue with automation tools is that we trust that their work is "good enough". But when it comes to beating your competitors, "good enough" may not be as good as them.

Nothing has or will ever be better than humans are creating effective content. Whatever the source.