Hummingbird & Entity Search – #SMXSydney
This next update from SMX Sydney is all about the Hummingbird algorithm update, which was launched in August 2013 on the eve of the company’s 15th anniversary, and which may fundamentally change the SERPs.
We’re introduced to this topic by Cyrus Shepard from Moz. Cyrus’ first point is to differentiate between Hummingbird and the Panda and Penguin updates. The Panda and penguin updates target spam and manipulative SEO., i.e. blackhat tactics (hence the black & white colouring?). Hummingbird however doesn’t target webspam. It’s about better query understanding, user intent and understanding what your content is about.
Panda and penguin are filters that sit on top of the algorithm. Hummingbird is an entire re-write.
There’s no ‘getting hit’ by the Hummingbird algorithm. It’s a different kind of algorithm, so it requires different tactics.Hummingbird is not entities, schema.org, synonyms, knowledge graph or Google Plus. However these are elements that are working together for improved query and content matching.
- conversational search
- better answer boxes (where answer is bolded, not they keywords)
- query rewriting (e.g. rackets to racquets)
- fewer long-tail results
There’s still a long way to go, these sorts of improved query matching so far are only affecting a small percentage of results.
Meaning becomes more important than individual keywords
Domain diversity is decreasing. That is to say the total number of domains ranking in the SERPs for a given keyword is reducing. This makes it harder for the smaller player. It’s no longer about that long-tail keyword, it’s about all the other ranking factors (which is where the big players win).
So what exactly is ‘Entity’?
It’s all about connections. How does Google understand exactly what the ‘entity’ is that is being referred to? e.g.
- Different keywords apply to the same entity: Barack Obama, President of USA. Michelle Obama’s husband, Commander in Chief
- Same keyword applies to multiple entities. Disambiguation. E.g. plant (as in a tree or a factory)
There are a few tools out there that help you to better understand those connections, moving away from traditional SEO keyword research tools. e.g.
The focus is on making connections with your content and data – and structuring it to improve Google’s understanding of it.
How do we make these connections?
Schema.org – it’s hard to implement but is worth it.
Not many people are doing it yet. But sites with schema.org show up in SERPs with much higher frequency. (NB. The caveat being that this could be because big sites are quicker to adopt schema).
According to Moz ranking factors, pages that use schema.org have only 0.02 correlation with higher rankings BUT Searchmetrics study found that sites with schema.org dd have a higher average ranking.
The difference? Moz’s research took place before Hummingbird was release while Searchmetrics’ took place after. Moz are about to repeat this research so we’ll be on the lookout for potentially updated correlations.
Tools to help implement schema:
- JSON-LD lets you send scheme.org markup using JSON instead of HTML
- WordPress plugins, e.g. Raven Schema Tool
- Validate with structured data testing tool
Schema is only one aspect of creating these connections. Google can figure connections out but needs guidance. We need to provide context, answer questions. Links still count. Google+ for Local provides clear entity connections. Much of this remains grassroots SEO.
Following on from Cyrus’ introduction to Hummingbird, Greg Boser follows up on its potential impact on mobile mobile strategy?.
There’s no drastic changes for how you approach mobile SEO. But it has created new window of opportunity. Hummingbird is a response to the dramatic increase in mobile search. This means Google is placing more importance on quality mobile experience. Big entrenched brands with poor mobile experiences will suffer, up and coming brands with quality mobile experiences will benefit.
Where to Start with Your Mobile Strategy?
- Forget about tech for now
- Understand behavioural differences between mobile and desktop users
- Understand this for your own site specifically (analytics, interaction)
- Determine which tech approach is right for your situation
So which technical approach should I use?
Responsive (Google’s recommended configuration)
Pros: completely client side, cost effective, easiest to maintain, single URL
Cons: Potential speed issues, inability to customise for mobile experience, potential for poor user experience for mobile visitors
What to remember:
- Properly optimising CSS/JS
- CSS/JS files must be crawlable
Pros: Reduces code bloat, greater control over user experience
Cons: Build and maintain multiple templates, Reliant on user-agent detection
What to remember:
- Use HTTP Vary header
Separate Mobile Site
Pros: Streamlined, optimised code. Able to custom tailor experience specifically for mobile
Cons: Maintaining two separate site, relies on user-agent detection and redirection, potential link dilution, lots of potential issues with misconfiguration
What to remember:
- User proper bidirectional annotations and redirection
Thanks to urbangardencasual.com for the hummingbird photo!