When it was first released we saw a massive surge in the SERPS that was unprecedented in our 20 year history yet as it was a retrospective announcement it explained what we saw happening a couple of months earlier.
This comprehensive slide set reveals just how significant this jump was both for our clients and for ourselves but the post in the end revealed something even far more important, the secret of why this happened
For nearly 18 months I’ve been trying to match up why Hummingbird caused this effect on our search results. Until last week
A simple diagram in the patent revealed the colourful little secret that I’d been looking for all that time. It revealed the exact cause of why our CISE semantic themesets are so successful in the SERPS and have retained and improved on that positioning in the long term.
In Google’s cleverly named Synonym Engine, I say that because synonyms are less than 25% of what is revealed that diagram, we see 5 clearly named functions. 4 out of the 5 have been clearly described online but one in particular has had absolutely no exposure whatsoever, i.e floating context.
So these five factors are:
Synonyms we get but who had thought about synonym pairs previously?
The manipulation of the query is most intriguing as it indicates the conversational sense we have seen discussed in many places..
Confidence values immediately rings a huge bell called Authority
And then we get to the real crux of the matter, context.
The first one covered is slightly more obvious i.e, adjacent context and this has always been a huge factor within CISE ever since it was envisioned around 12 years ago even so we created these quite unknowingly in the initial stages of our testing.
BUT .. undoubtedly the real game changer is floating context.
Let me explain. Adjacent context is quite obvious like web development is to web design even down to the hierarchical elements such as techniques and standards are and just like SEO is to web marketing or as analytics is to web marketing but what in essence is this floating context and how do we harness it ?
Floating context, in my view, describes contextually related words and phrases that do not necessarily have to be in the same context of the content being examined. That’s why a seemingly non-related term may even help rankings especially from a creative viewpoint. CISE produces these type of terms through morphological patterns. It’s quite obvious when you realise it, it’s exactly how comedians get their jokes simply by loosely playing on the context.
Recently I wrote an article about body language in web design. I got this idiom directly from a CISE search on web design. It sort of made sense as there’s <body> tags in the code and what is in between that is language but at best it’s floating context but on the other side of the coin it’s a strong creative metaphor so it makes the content not only unique but metaphorically attractive. For me it was the perfect vehicle to explain how semantic optimisation works. See the article here
So over the past 18 months we have tracked these obscure words and found they helped our ranking substantially but until last week we didn’t know exactly why that happened.
So if you read this correctly, it opens up a huge door of creativity where apparently loose context is treated as if it was a semantic keyword. So when we infuse our content building up entities, tuples and triples we are feeding the googlebot with exactly what it’s looking for. It’s what we term Semantic Infusion.
CISE creates these semantic themes sets through a complex algorithm that considers a range of language patterns and there’s no mind on earth that can even get close to the highly useful lateral suggestions it spits out. But even if you don’t use CISE you can walk away knowing that the boundaries are far wider than you thought they were, but guessing what these boundaries actually are aren’t as easy without it.
So my advice is get creative, buy some CISE themesets or just get in the zone because these are the sort of sites that will really make the grade in the brave new semantic web