Two of the biggest SEO news stories of recent years concerned the Penguin and Panda updates to Google’s algorithms, which were aimed at improving search engine results by combating spammy black hat techniques and returning higher quality results at the top of the SERPs. While these changes had a huge impact on SEO, they were just tweaks and modifications of Google’s existing algorithms. It appears that Hummingbird, which was silently implemented a few weeks ago, is a much more technically significant, perhaps the most thoroughgoing update since Caffeine in 2010 or even since the algorithms were significantly rewritten by Amit Singhal in 2001.
Hummingbird is estimated to affect 90% of search engine results and is likened by Search Engine Land’s Danny Sullivan to a complete replacement of the engine of search.
Hummingbird is part of Google’s ongoing transition from simple keyword-based search to the promotion of a more conversational mode of interaction. Keywords are limited because they fail to account for user intent and multiple meanings (polysemy). A lot of information can be brought to bear in order to figure out user intent, including location data, social media data, semantic analysis, and the relationship between queries.
For example, using keyword-based search, “I need a new iPhone” might bring up a list of search results that include news about a new iPhone release, the Apple site, articles about apps to install on a new iPhone, reviews, Apple rumor sites, “what you need to know about the new iPhone” articles, and so on. These results fundamentally fail to understand the intent of the searcher and forces them to rephrase their search or trawl through irrelevant results to find what they need.
A more conversational approach would understand the “meaning” of the sentence and from the meaning figure out the intent of the searcher, enabling Google to quickly return exactly what the searcher needs.
Hummingbird is intended to improve Google’s search results for longer and more conversational search terms that reflect how people would naturally express their desires. It’s part of Google’s move towards Semantic Search as reflected by the Knowledge Graph and especially by Google Now, their mobile search application.
As usual, Google insists that these algorithm changes don’t impact on their advice for webmasters and SEOs. Google still wants us to create high-quality, informative content and leave the details of how that content finds its way into the SERPs to them. But, although Google is not giving out any technical details, it is possible to speculate about some SEO best practices that should be implemented to take advantage of semantic and conversational search:
Schema.org metadata — However smart Google’s algorithms have become, they still need some help identifying the entities that occur in text, including locations, products, individuals, and the relationships between them. If Google intends to discover intent from conversational queries and reply accordingly, those sites that make it easy for Google to determine the meaning inherent in their content with metadata are likely to be at an advantage.
Long-tail keywords — Conversational queries are inherently idiosyncratic, and although Google is probably using complex substitution rules for determining the relationship between semantically related clusters of keywords, it makes sense to give them a helping hand by using a diverse set of keywords, rather than focusing exclusively on a small set.
Ensure that your content aligns with searcher intent — In many ways, this is what Google means by high-quality content. Thin content and other low-quality content that previously worked well in the SERPs because of Google’s keyword-based pattern-matching strategy generally doesn’t have any relationship to user intent. Businesses and marketers that understand their users can create valuable content that matches with buyer’s needs.
In short, if you’re already doing the right thing, nothing much has changed from the perspective of webmasters.