We've been through this before. Sometimes the hype is understated, as when Penguin rolled out and at first it did not seem like a big deal. Now it's clear Penguin is a very big deal with many severe penalties for naughty link-building. Panda, in contrast, doesn't seem to be nearly the problem.
Hummingbird and Latent Semantic Indexing (LSI)
Hummingbird seems to be mainly about latent semantic indexing, which is a fancy word for:
- Google being smarter about what words go with what words, and thereby to infer meaning. So many words and phrases are ambiguous and only with "context" can a "machine" interpret what you truly mean.
- Moving to a more Siri-like user experience. Allowing users to "talk" to their computers and get responses.
Do you know the best Pizza restaurant in Palo Alto?
The relevant words are: best, pizza, restaurant, Palo Alto.
No amount of Hummingbird technology can "change" the fact that speech queries are driven by keywords. At its best Hummingbird will be about inferences, such as:
Do you know a pizza restaurant?
With the inference being you want a) the best, b) someone near you, and c) I know you are in Palo Alto.
Is Hummingbird important? It's too early to tell. Can Google or Will Google change the fundamentals of human speech and queries? No: at best they are going to try to be more intelligent about adapting to the ambiguities of human speech, and better at inferring what a query means.
From the SEO perspective, it all still implies that we should:
- Know our keywords.
- Write strong keyword-heavy copy, that is good both for humans and for Google.
- Build links, social mentions, and all that external stuff that confirms we are an "expert" on the topic.