Last updated: What the Google Hummingbird algorithm means for e-commerce

What the Google Hummingbird algorithm means for e-commerce


Listen to article

Download audio as MP3

We wanted to know what the new Google Hummingbird algorithm means for e-commerce.

So, we asked Grant Simmons, director of SEO and social product from The Search Agency, the largest independent search marketing firm in the United States, to share his perspective on just what retailers and B2B marketers need to know in order to hang onto those oh-so-valulable search rankings. Here’s what he had to say.

First, can you explain exactly what Google’s Hummingbird algorithm means, in layman’s terms?

The new algorithm is Google’s biggest change in search curation over the last decade where its algorithm has been updated to provide better answers – rather than just search results – to search queries. Hummingbird allows Google to better analyze your whole search query, rather than just parsing word searches, and provides better answers based on your intent and semantics of your question. Google is also leveraging additional data points to draw better insights into your query based on several signals, such as your location, device, social connections, and previous searches.

What kind of impact has Hummingbird had so far, and what impact will we see down the line?

At this point data is pretty thin, as demonstrated by Google’s announcement being more of a surprise to many, as opposed to more noticeable impact of past Google updates such as Panda & Penguin. Both these updates focused on the ‘now’ of site tactics, looking at content and linking in a traditional SEO lens.

The challenge with Hummingbird is it goes beyond just an overlay of existing algorithmic signals, but rather changes both the interpretation of a user’s search query and the methodology of how those match to web content. As Google tweaks Hummingbird to iterate the relevance of matches we’d anticipate that more complex queries (more words or conversational queries) replace the one or two word queries SEO folks are used to optimizing for targeted traffic. Example: “best family car” is replace by “where can I buy the best car for a family of 4 near my office?”

What does Google’s Hummingbird algorithm mean for retailers and B2B e-commerce, and what strategies do they need in place to take advantage of the changes, or stay on top of their SEO to maintain their current search rankings?

I’d recommend first an audit of the metrics retailers already have access to. With the disappearance of keyword level data in organic search (at least from Google Analytics) via keyword (not provided), SEO Marketers need to think beyond the keyword and really dive down into what user *behavior* is driving interaction, engagement and transactions.

Both Google Analytcis and Omniture provide page level metrics, but these should be viewed through key “themes” rather than traditional keyword lists, so analysts can demonstrate how pages answer specific queries that align with user needs and company products, services and activities. i.e. “where can I plug my washing machine in?

Google Webmaster Tools provides also some metrics, but more importantly trends, technical issues and technical recommendations to improve search engine “friendliness.” (Bing also has an excellent suite of Webmaster Tools that can help webmasters align with search engine (and user) best practices.)

All this data provides insights in the post click site activity, but retailers and B2B e-commerce sites also need to better understand user behavior pre-search. Online channels offer massive opportunity to influence buying behavior, so it’s important to look at potential customers as individuals, with individual needs, buying patterns and preconceived behaviors. Segmentation and alignment of intent, context, devices and channels is key to providing a better experience which will both endear brands to their customers as well as attract new users into the fold.

In addition to better target marketing, fundamentals of search need to be maintained. Google is the most discriminating of users, so webmasters, marketing departments and SEO Marketers need to work hand-in-hand to ensure the user experience is close to flawless. Social media, PR, paid media and IT professionals also need to be coordinate, to ensure messaging aligns with platforms, interaction, technology and device.

You’ve said in the past that Google is further underscoring the importance of user intent over pure keyword based interpretation, with the added layers of context, recency, conversational search (demonstrated in May this year), location cues, Knowledge Graph data and better understanding of complex queries. Can you talk about how that differs from the “dumb keyword” approach to SEO?

This is the really key concept to better relevance matching. We often talk about intent to content, but there’s a lot more data that search engines can leverage to fine tune how they curate their search results. Context is derived from a lot of those data points, things like location, device and same-session searches, but there’s certainly more signals from social, business or behavior graphs that are probably part of this fine tuning too.

A good scenario might be a search query on a mobile device for ‘pizza’ where Google will curate results to show only pizza stores en route to your home, only those that are open, stores you’ve previously purchased from, and businesses where you’ve given a positive rating on Google+. A little scary perhaps, but privacy concerns aside, wouldn’t the average user prefer a search results experience that only presents the most relevant of results?

I think that the big surprise in search engine innovation is that it’s relatively transparent to the user. “Dumb Keywords” weren’t really due to dumber users, but the user tolerance for dumb matching made the results bearable. Now empowered users are asking their phones complex questions via voice, and are demanding smarter answers, and I think this is more of a feature of great marketing, proliferation of smarter technology and the subconscious thought of it “just works”, that drives user adoption and interaction.

From an SEO standpoint the obvious challenge is moving ‘dumber’ processes to more user-centric practices. Queries can be ambiguous for person-to-person conversations, so the difficultly of both Google and SEO folk ‘best-guessing’ user intent and context shouldn’t be underestimated. There are, however, key opportunities. In the same way Google leverages many data points, so SEO practitioners should too. Intent and context behavior have some predictability, which can be leveraged for planning, forecasting and help define relevant ‘metrics that matter’ for SEO justification. But traditionally these data points, such as user personas, demographic segmentation and upstream / downstream tracking, have been the purview of paid search or marketing departments.

Therein lies the solution. SEO should be considered as and treated as a true marketing discipline with as much focus on the technical side as the marketing strategy. Understanding users *before* they interact gives SEO Marketers better data-driven cues as to *how* best to connect with their most valuable audiences, moving SEO from dumb keyword matches into the realms of query-focused analysis, solutions-based content development, targeted platforms, segmented tactics and better tracking of key metrics.

A lot has been said about Google’s Authorship push that has authors connecting articles to publishers to author profiles which in turn creates signals of trust (or mistrust) by association and provides a topic alignment with author expertise. This will certainly lead to trust signals for more granular expertise too – like recommendations on the best family car rated by the most authoritative automobile experts. These evolving signals of trust, relevance, expertise and strength of association will be important as Google seeks to return results that are definitively “expert”, through the intelligent presentation of knowledge-based results.

In your opinion, how will search continue to evolve in the future?

There are 3 main areas where I see search evolving, all of which are currently available, but not necessarily perfected nor adopted:

  1. Voice search. Sure we have Siri and rudimentary voice interfaces, but Hummingbird, processor power and software innovation is making the ‘perfect’ voice search interface a closer reality.
  2. Hyperlocal offers. Just like “Minority Report’ where Tom Cruise is recognized in a Gap store of the future and offers personalized specials in his size via a large video sales pitch, so hyper localization can drive sales at relevant connection points in day-to-day routine, like food, services and / or real life product placement. I predicted in an article a couple of years ago the fact that our cars will be making relevant ‘passive search’ recommendations to tie need to location. It’s happening. Has your car talked to you today?
  3. Wearable search interfaces. Google Glass is the most apparent technology that ‘fits’ the bill, but smarter watches, jewelry, clothing and less intrusive devices will revolutionize search through visual, audio or touch based prompts. Nike bracelets, video label pins and other currently-in-development platforms, driven by gesture or even biometric signals can give better understanding of context, sentiment and activity, to assist in passive search answers e.g. as you sweat on your run, a voice suggests you cool down with a Coke at the next convenience store .2 miles down the road.

Last thought: Retailers and B2B e-commerce *are* evolving. Savvy merchants and ancillary services will evolve and profit. Understanding how Google is changing the landscape and is iteratively changing how people shop will be fundamental to traditional retailer survival. Although “adapt or die” is pretty harsh, it *will* be the innovators, the adaptable and the flexible who will grab market share and thrive.

No cookies? No problem.
(Unless your marketing platform is rooted in The Old Times.
Then you’re in trouble.)
Attract, convert, retain: It starts HERE.

Share this article


Search by Topic beginning with