So what exactly is Latent Semantic Indexing? When asked this question, I typically struggle to come up with an answer that’s simple to understand. So I thought I’d have a go at laying it all out in a blog post, rather than confuse any more innocents with a verbal answer made up on the spot.
Latent Semantic Indexing (LSI) is a component of a search engine algorithm that focuses on trying to understand the meaning of the words on a webpage. Not the dictionary meaning of the words, but the intent behind them.
This isn’t anything new. Google has made its reputation on delivering results based on what you mean, rather than on the exact words you type. Take this search, for example: “Black and white movie with Billy Bob Thornton”.
This is a perfect example of LSI in action. Google’s results have looked past the words I’ve used and found the intent behind them, in this case delivering information about the Coen Brothers film “The Man Who Wasn’t There”.
Interesting, but from a perspective of SEO usefulness, there really isn’t much you can do to turn LSI to your website’s advantage.
To work through a practical example, let’s say you own the website http://www.awesomelocksmiths.com.au , and one of your pages talks about security locks. This page would be hard pressed to talk about locks without also using words like “key”, “pin” or “security”. LSI doesn’t just rely on the actual keywords, but the rest of the page to determine which kind of “lock” you’re talking about.
If you’re thinking that this all sounds needlessly complicated, remember that dictionary definition of “lock” isn’t that simple. In fact, the word “lock” has 8 definitions:
- A mechanism for keeping a door, lid, etc., fastened, typically operated only by a key of a particular form (e.g. the key turned firmly in the lock)
- A similar device used to prevent the operation or movement of a vehicle or other machine (e.g. a bicycle lock)
- (in wrestling and martial arts) A hold that prevents an opponent from moving a limb
- A number of interlocked or jammed items (e.g. a street closed by a lock of carriages)
- A short confined section of a canal or other waterway in which the water level can be changed by the use of gates and sluices, used for raising and lowering vessels between two gate
- An airlock
- A person or thing that is certain to succeed; a certainty
A mechanism for exploding the charge of a gun
So how does Google figure out which “lock” you’re talking about? It decides based on the semantics of the page content, the context of the search query, whether your searching on mobile or a desktop, and probably a thousand other factors.
Meaning that if I search the term “Locks Sydney”, Google has to determine if I’m searching for “Security Locks” “Locks of Hair”, “Canal Locks”, etc. You may think that adding a location keyword would help Google understand that I’m in fact searching for a locksmith in that area, right? But take a look at the results for “Locks Sydney”:
Now granted, If I did need a locksmith in Sydney, then I probably wouldn’t use that exact search, but it does demonstrate LSI in action. Markus Schulz locks in Sydney NYE Show is on the first page of the results. Which is interesting because Google clearly knows I’m looking for a locksmith in Sydney. The related searches are entirely about Sydney locksmith companies. Even after LSI has been applied, Domain Authority still has an overriding effect on the results!
Well, nobody has ever accused Google of being easy to understand. That would take the fun and challenge right out of it. Nonetheless, I hope that this post has helped shed a little light on the mystery of Latent Semantic Indexing – just in case you were wondering.
Any questions? Let me know in the comments.
- 7 Simple Principles of Website Design That Impact User Experience - August 3, 2021
- Apple’s privacy update (what does it mean for small businesses)? - July 15, 2021
- 6 Influential Ways Web Development Impacts Your Marketing Strategy - July 8, 2021