A few years ago terms like “LSI” (latent semantic index) were ubiquitous, which is pretty much the pattern of every “new” thing in marketing. This little 3-letter nugget speaks to a mathematical method of grouping related words together, which search engines can use to expand the usefulness of results to queries.

For instance, some people refer to watches as “time pieces”. A site seen as an authority for high end watches naturally should also rank for time pieces, and as Google became sophisticated enough to associate these words that indeed began happening.

But folks can go overboard trying to use LSI, study it, and ultimately invest more time in it than is beneficial.

“What! But there are SEO blogs all over the internet harping on how major LSI is! How am I wasting my time?”

I’ll explain.

Latent Semantic Indexing Is Machine Learning

Yes, as SEOs we’re constantly trying to anticipate and understand Google’s algorithm so we can give it what it’s looking for. So in the strictest sense, trying to understand how it’s grouping related words is beneficial.

However, how this is accessible to the average user is usually pretty obvious. In our example above, you wouldn’t need any kind of in-depth research to know that “time piece” is another name for a wrist watch.

The reason LSI seems so complex is because teaching a machine to think like a person doing a search is extremely complex.

That complexity causes some people to go shoulders-deep into linguistic psychology, hungry to find all sorts of hidden word associations no one else has considered and make them a part of their SEO campaign.

But there’s a hangup in this reasoning: Google is using LSI to find the most common yet different ways searchers look for answers. Even if you could find some obscure ways to rephrase search queries, would they even be useful?

The Point Is To Focus On Similar Ideas

You’ve heard of the Law of Diminishing Returns, right? The idea is that something can be useful at first, but past a certain point having more of it becomes less and less useful.

Studying the psychology of word choice is useful, but more so for copywriting than for SEO, in my experience. You can improve someone’s SEO significantly simply by telling them to keep similar and related keyword concepts in mind when writing page material.

Doing a bunch of psychological research past that point might be mildly helpful, but when weighed against the time investment tends not to be practical for the average business owner.

  • By avoiding a number of extremely similar pages to cover each variation of a search idea (like blue watch, blue time piece, blue watch band, etc.) your average quality score will improve (which builds authority).
  • You can do more with less, and invest less time building everything.

Put simply, if common sense allows you to “get it right” a lot of the time, that’s usually good enough. Knowing to even think in terms of keyword variation is most of the battle — e.g. don’t overuse the same keyword over and over, but vary it.

And if you are stumped for ideas on other ways of describing your products and services, I find it’s simpler to ask your sales reps what prospects ask about, or ask your customers directly.

Find this useful? Check these out, too: