Thanks to this week’s guest post from +RickEliason from the Reload London office.
To quote a colleague: Hummingbird changes everything! Business owners, marketers and agencies should be revisiting their keyword lists to ensure they now comply with user intent – out with the generic, high-traffic keywords and in with the intent-focused keywords. That’s the message with Google Hummingbird!
Ok, slow down a little. Hummingbird is still in the early stages of deployment since its quiet release (around a month ago) but already, there is talk that around 90% of search queries are affected. Google’s search chief Amit Singhal himself stated that Hummingbird isn’t just an update to their existing algorithm like Penguin and Panda, it is a complete reworking. This is the direction of search now so take heed of this article and adapt accordingly.
So, what is Google Hummingbird?
Hummingbird is Google’s latest offering in the battle to return the highest quality search results possible. Google have adapted their new algorithm to not only accommodate changes in user search behaviour (particularly with the advancement of mobile technology) but also further utilise its own understanding of the WWW. This update was launched several weeks ago with very little PR and coverage and has only recently been confirmed by Google themselves. This update marks a new era of SEO and online marketing.
How Does Hummingbird Change Things?
Hummingbird is a leap forward in understanding user intent (going beyond just typed search queries). Thanks to conversational search (voice powered), queries are becoming longer and more specific and Google is better understanding the ‘intent’ signals behind searches.
Before, a search for [Botox] could yield results for medical practices, information on the injectable as well as news stories on celebrities using Botox. Hummingbird aims to use certain signals to filter out irrelevant results so moving forward, a site shouldn’t rank for both [brisbane botox clinic] and [celebrity botox disasters] when optimised for [botox]. Likewise, a search for [botox] by an individual should not yield celeb disaster articles alongside clinic review websites as Google uses implicit ‘intent’ information to try and decipher what type of results the individual is after. So what are these ‘intent’ signals?
Types of Intent Signals…
First of all, please watch this Whiteboard Friday video on implicit vs explicit queries by Will Critchlow of Distilled. Don’t worry, I’ll wait.
Mind-blowing right? Ok, let’s carry on. Intent signals is all of the implicit information Google knows about you right before you enter your query such as the following:
- Who is searching? – Based on your previous searches and behaviour are you male or female? Are you a teen or pensioner? Are you more interested in daily gossip news or do you tend to search around a particular hobby? Based on your Google+ interaction what are your interests? Who do you follow? Based on your Gmail inbox are you a bargain shopper or an avid traveller?
- How are you searching? – Are you on the move with a mobile device or are you at home on a laptop? Is the query a typed set of keywords or a more conversational spoken query using voice activated search?
- When are you searching? – What time of day/day of the week is it? What have you been searching for in the last 10 minutes?
- Where are you? – Where are you right now? Where have you been in the past?
- What type of search are you performing? Navigational (do you know the brand you are after?). Informational (do you need specific facts or info?). Commercial (are you shopping around). Transactional (are you ready to buy?)
This is by no means a complete list. Google uses intent and implicit information from all manner of sources including Gmail, search history, cookies, user profiling, +1s, Google+ usage and connections and more in order to return the best results.
How will Hummingbird affect businesses?
Less Traffic: Intent usage is already in evidence with knowledge graph results, blended results, auto-type, spelling corrections, etc. Each of these is potentially directing traffic away from your site and towards other resources meaning less traffic. Google’s Knowledge Graph for example keeps the user on Google properties as long as possible. That way it can serve authenticated and trusted results to users. Failing that, it has a growing database of trusted experts on Google+ to fall back on – If you are not using Google+ to your advantage, you could be losing out on traffic now.
Less Keyword Data in Analytics: As Google weighs keyword queries along with implicit signals, the keywords themselves become less relevant as a traffic source indication. For instance, a restaurant website might receive traffic for the keyword [breakfast in brisbane] not because it is optimised for such a query but because the query was performed on a mobile device in the vicinity of the restaurant at breakfast time and several of the user’s Google+ connections had recommended the restaurant previously – therefore Google is hiding this potentially misleading data altogether.
More Targeted Traffic: Or at least for those that appreciate these new developments and adopt a new keyword targeting strategy away from top-level generic terms and towards long-tail, hyper-relevant terms. The good news is that with Google understanding sites better and the intent behind searches, there will likely be less competition for the sites doing it right.
How to do it right
- Marketers should revisit their keyword list and (generally) ditch the broad, high traffic keywords in favour of those super-relevant to the website’s reason for existence
- Lower traffic figures per keyword need to be communicated to clients/managers at keyword selection stage
- Appreciate the growing usage of spoken queries on smart phones and factor this into content
- Learn to track and use internal site search data to reveal intent terms and low-level keyword data
- Understand that keyworded content for ranking purposes doesn’t have the value it used to. Content must have real-world value and serve a real purpose.
- Freebase & Schema go together to feed the knowledge graph http://youtu.be/5lCSDOuqv1A