hummingbird-Illustration-WEB

Hummingbird: Move Over Caffeine, Hello Sweet Nectar

A quick glance at Twitter or your RSS reader will reveal that everyone is all a’flutter about Hummingbird, Google’s latest algorithm update. This release is the biggest one since Caffeine in 2010. How is it different from its predecessor?  Amit Singhal reports:

“[Hummingbird] is clearly more focused on ranking sites better for relevance, while Caffeine was more focused on better indexing and crawling of sites.”

Undoubtedly, this is big news for the industry, but it’s not yet completely clear what will change. To get some insights, we caught up with Bill Slawski at SEO by the Sea, resident patent expert:

Bill-Slawski

First, can you tell us what Google patents are being used by Hummingbird?

I don’t know for certain if I know the patents behind Hummingbird, but my intuition tells me that they are significantly related to better understand the contexts and syntax of queries. I wrote blog posts about two different patents that appear to be related, and I’ve been working on another for a couple of days that seems possibly even more related. My two posts are:
1. How Google May Reform Queries Based on Co-Occurrence in Query Sessions looking at the patent Generalized Edit Distance for Queries
2. Relationships between Search Entities, which looks at the patent Search entity transition matrix and applications of the transition matrix

The third patent is Synonym identification based on co-occurring terms, posted on today: The Google Hummingbird Patent?

Do you think schema and structured data will play a role?

I don’t see a role for schema or structure data at all in this process.

In your opinion, how do you think Hummingbird will impact SEO?

Hummingbird will enable people to create longer, more complex, and more natural language type queries, and receive higher quality results, even if all the words used in the query don’t appear upon the pages being returned.

That last question – how Hummingbird may impact SEO – is one that we were particularly interested in, since it involves a combination of expert opinion and a bit of good old-fashioned guesswork. We wanted to hear what others had to say on the topic as well, and so we pestered a few more people who graciously answered (even after business hours – thanks, folks!)

 


Ronell-Smith-2

Ronell Smith

As a content strategist with an affinity for SEO, I couldn’t be happier Hummingbird arrived. To my mind, the algo change will lessen two of my biggest frustrations.

First, I’d come to hate Google for sentence-baced search since the algo was only picking up a few of the words in the string. For example, a while back I typed “How to fix iMessage problems when switching from Apple to Android?” The answers I got back centered around buying a phone, either an iPhone or a Samsung.

Not much help when you’re just trying to get clarity on how to stop the iPhone from tagging a number to iMessage. With the Humminbird, this should not be a problem, since Google says each word in a query will receive attention.

The second reason I’m happy about Hummingbird is I think it’s going to reward content strategists and content marketers who do their homework. I’ve long railed against content marketers who focus more on the “marketing” component as opposed to being willing to get in the weeds with the strategy.

These folks are able to create lots of poorly guided content, much of which does nothing to sate the appetite of those they are marketing to. I refer to these folks as creating messages meant for no one in particular.

There is no attention given to personas or linguistic cues, and context is ignored. The pre-Hummingbird algorithm still rewarded these folks if they picked the right keywords.

Now, however, those who nail strategy when content-planning will be richly rewarded. Since they’ve likely parsed out subtle cues in the language, and because they have a great idea of who they are marketing to, the sentences and phrases they type into Google should yield a depth of context-rich information that can further guide the process. I’m going to enjoy Hummingbird, even if it’s not a be-all and end-all.

 


Beth-AndersonBeth Anderson 

I think – and this is entirely conjecture – that it may result in a little bit more churn when it comes to SERP results. Even though the engine is named as such to evoke speed and precision, the information that’s come out is focused on conversational search, the kinds of things I do when I’m walking around Seattle and want to find “restaurants near me” (which incidentally got a great result for me last night! thanks Google!).

If you mix conversational search with the ability to be fast and precise, I suspect that it means that the search engine itself will be a little more responsive, and learn a little quicker through user behavior which results people are actually looking for, which should enhance its ability to be precise. I’m not sure how personalized that result might be – whether if I search for “restaurants near me” Google will know that because I’ve searched for Vietnamese restaurants in the past, it will present me first with the pho restaurants in a six-block radius, followed by other results, but we’ll see.

In terms of how that affects SEO – I think the key is to continue to have relevant, engaging information for customers, which does include continuing to build out localized and long-tail content – if the search is going to try to get better at answering conversational questions, our sites should have those answers somewhere, preferably somewhere easy to get to and from.

 


Jayson-DeMers Jayson DeMers 

Hummingbird is a clear step toward improving the intelligence of Google with regard to more complex, conversational search queries.

With this change, Google is anticipating an accelerated rise of mobile device usage, and specifically more voice-input search queries. Voice-input search queries tend to be longer and more colloquial in nature than typed search queries, so Google knows it needs to get intelligent enough to accurately handle the queries.

For the SEO industry, this is a big signal that mobile optimization, and specifically a mobile content strategy, is going to be necessary for getting visibility in Google over the next several years.

 


Joel-KlettkeJoel Klettke (author of Hawk Attack)

Oh boy! Well, here’s what we know about Hummingbird so far: NOTHING. Or rather, that it is “Specially designed to handle complex queries”. It’s all so clear! So, now it’s time to do what I imagine all the thought leaders are doing: Make wild guesses associated with junk stats!

Here’s what I think: A long, long time ago, Sergey Brin told an interviewer that he wanted to get to the point that people didn’t even need to type in a query to get the information they wanted. No. Queries. That’s the future of the internet.

Now, let’s assemble what else we know:

 

1. Google is putting a huge emphasis on Knowledge Graph and entities

2. As seen in Dr. Pete’s Beyond 10 Blue Links presentation, Google is also acting like a massive aggregator, trying to keep people from ever having to leave their shiny, white interface.

3. Google just took a whole bunch (namely, all of it) of organic keyword data hostage and told everyone it was for privacy. Then they went out and shot some photos of every street in the known universe and read all your e-mails. INTEGRITY!

4.  Google continues to force everyone with a pulse to get onto Google+, going so far as to trick them into signing up.

5. Amit Singhal keeps talking about how he wants to build a computer like the one on Star Trek that answers any question he asks it and understands logical question strings (Like: “Where is the Eiffel tower”, followed by “How tall is it?”)

So, if we throw all these things in my speculative blender and title the outcome “What Joel Klettke Thinks of Hummingbird,” we get this:

Hummingbird is likely Google’s attempt to leverage the Knowledge Graph across a broader array of queries. I wouldn’t be shocked if they were interconnecting queries and search history to try and guess at the intent behind a search, following logical strings. I think when we think about “complex queries”, we think about people with specific questions. Like comparisons, which Google also showcased today – a side-by-side comparison box in the SERPs. I’d imagine Hummingbird is their means of trying to answer more complex questions with Knowledge Graph and the data they already have.

I’m guessing that KG is getting better and better at ripping apart elements of entities and using those to answer questions in helpful little boxes that rob every other site in the world of revenue.

Ta-da! But don’t take my word for it! Seriously, don’t. I don’t know anything, and anyone who says they do is trying to earn your clicks, which they will then cash in for Chuckee Cheese tokens and speaking gigs. (hint, hint. Speaking gigs. Joel Klettke. Logical string.)

 


Pete-MeyersDr. Pete Meyers

It’s still very early, but Google is suggesting that “Hummingbird” is on the scale of Caffeine and may be a full-scale change to how the algorithm works.The Caffeine connection also implies new infrastructure and capabilities. When Caffeine rolled out, SEOs just heard “techie stuff blah blah blah”, didn’t see an immediate impact, and went on with their lives.

The reality is that Caffeine powered major algo updates for months, including (probably) Panda and Penguin.

Hummingbird could be the building blocks for the next stages of semantic search and the next major algo updates – we could be seeing the impact for months or years.

 


Travis-WrightTravis Wright

Google says that the Hummingbird update is impacting greater than 90% of all searches, and that it was implemented a few weeks ago on the down-low.  They’ve not shared too much information about this update, other than it was the biggest update to the Google search algorithm since the Caffeine update back in 2009.  Page also mentioned a tweak to the Knowledge Graph, which focuses on parsing searches to handle comparison questions.  That could come in handy.

What I’m most upset about in the SEO world at this time, isn’t Hummingbird, but the fact that all search traffic from Google in Google Analytics is showing up as “100% Not Provided” on all non-Adwords searches.

I call that the “Dodo Bird” update.  This will really make the job of doing SEO a bit more difficult.. but that always seems to be Google’s modus operandi, doesn’t it?

It’s finally time to embrace structured data and semantic SEO and implement it across your sites.  

 


Anthony-PensabeneAnthony Pensabene

Since, quoting Google stormtroopers, the algo tweak is addressed toward more complex queries, I’ll venture to say, marketers may have a better chance to tinker with longer tail keywords and phrases as well as consider how a given product/service lends itself to other entities.

To get more specific with that last part, I see the algorithm’s evolution like that of a child:  As the child has more experiences in the world, more discussions, and becomes exposed to more scholastic and practical information, the potential for more in-depth conversation with them increases.

With G mentioning they’re receiving ‘more complex’ queries (due to voice search), the way people leverage a query is evolving.  Through typing, we kind of talk to Google like cave people, using very broad concepts (kind of like a small child).  But waging voice search, we’re treating Google more mature, and being now 15, it’s had the time to grow it’s Knowledge Graph; it can have a more in-depth conversation, or so Google hopes.

All that being said, marketers may benefit from using such tools as Ubersuggest, minding what other kinds of information is related to a given keyword or phrase; it seems the more intelligent and in-depth ‘conversations’ brands index, the more chances they’ll have in Google aligning their information with users chasing specific wormholes.

 

Kate Gramlich Roumbos

Just a cat lady tackling the wild world(s) of SEO and content marketing.
Find me on twitter @kategramlichseo and G+

Get Free Updates
Given email address is already subscribed, thank you!
Oops. Something went wrong. Please try again later.
Please provide a valid email address.
Thank you, your sign-up request was successful! Please check your e-mail inbox.
Please complete the CAPTCHA.

32 Responses to “Hummingbird: Move Over Caffeine, Hello Sweet Nectar”

  1. A.J. Ghergich A.J. Ghergich said ...

    I really need to get around to reading Hawk Attack!

  2. anonymous coward anonymous coward said ...

    i just want to know why hummingbird, panda, penguin – why are these algo updates all named after exotic animals

  3. Rich Benci Rich Benci said ...

    Thanks for pulling together these expert opinions, very helpful! Although I do disagree with the patent expert, I think these changes ABSOLUTELY makes Schema.org markup much more meaningful, and that structured markup is what allows Google to verify and authentic web page information in order to understand that info in context with the search query … resulting in increased relevancy of the search result. #Algebrize

    1. A.J. Ghergich A.J. Ghergich said ...

      Hi Rich,
      I was kind of assuming the same thing about schema and structured data. It seems on the surface that would make it easier for Google to pull and classify your data…but I certainly defer to Bill on these things. Maybe Bill can elaborate a little more on this?

      1. Amy smith Amy smith said ...

        In my view Hummingbird comes on the heels of Google’s efforts to encourage sites to include semantic mark up-specifically schema.org mark up in content to identify data more thoroughly and create connections to wider concept.We should use semantic markup to our sites or make use of Google webmaster tool “data highlighter” as a temporary alternative. unless we will be steps behind those who are and gap will growing fast.

  4. Terea Jennings Terea Jennings said ...

    Thanks for clearing everything up. Nice try. Fortunately, I love a mystery!

  5. Bill Slawski Bill Slawski said ...

    Hi Rich,

    Thanks for asking. Google definitely likes the use of schema and structured data because it can make it easier for them to crawl pages and index the content found on those pages. It’s not completely essential in many cases, but in using it can result in rich snippets and better organized content at Google.

    Google has been awarded a number of patents and produced a number of white papers that shows them taking advantage of structure and semantic markup within HTML to do things like understand the semantics of content found within lists, which Google used for years in Google Sets. Google has used the semantic markup of HTML within tables, when used for tabular data as opposed to page and content layout, to do things like their WebTables project, and the technology behind that has been used in the past for Google Squared and to generate query refinements. Google has given pages breadcrumb rich snippets in search results when breadcrumb navigation has been used, even without the meta data/semantic markup that might make it easier for Google to generate those. Google Maps has been extracting location/address information across the Web without hcard or schema data that might make it easier for them to do so.

    Google has been granted at least 4 patents lately that look at a different kind of data that could be used to reformulate and expand queries where no mention of schema or structured data was mentioned nor alluded to in any want. Hummingbird appears to be focused primarily upon how a query might be changed, amended, and understood within the context of all terms within that particular query.

    Most of the news that we’ve received from Google about Hummingbird focuses upon Google better understanding most of the words within a query, and how those might be reformulated to focus upon better understanding the possible or probable meaning of words within queries with as high a level of confidence as possible.

    It appears to be aimed at longer and more complex queries, most likely for spoken or natural language queries that people might often speak rather than write. It seems that when people type a query, they use a minimum amount of words, but when they speak a query they abandon limiting themselves to just what they might feel are the best keywords and will ask full sentences, filled with pronouns and words that might not necessarily appear on the pages that they might be looking for.

    An example might be something like, “tell me about a place where I can get great chili and microbrewery beer.” A searcher isn’t likely looking for a page that calls itself a “place,” on it’s website, and Google might try to interpret the meaning of the word “place” in that query and replace it with something like “restaurant.” That would be one candidate synonym or substitute terms that Google might explore. Another might be “store” or even “street vendor.” Google may then look at the other words in that query, and decide that it should ignore some of the “skip” words within the query that may not add much value and see how well the each of the candidate synonyms or substitute words that could replace “place” might fit with other important and meaningful words within the query.

    It might look at similar queries within query sessions for when people searched for a restaurant to see if “restaurant” and “chili” showed up together. It might look for queries when “restaurant and “microbrewery beer” showed up together. It may do the same with the other potential candidate synonyms or substitutes showed up in queries within query sessions. Did people search for “store” and “chili”, and “store” and “microbrewery beer”? It may do the same with “street vendor” and “chili” and “street vendor” and “Microbrewery beer”.

    In this analysis, Google may be looking solely at information from their query logs, and determine that their candidate replacement term of “restaurant” tended to show up much more frequently with those other terms (“chili” and “microbrewery beer”) in query sessions, than “store” and those two other terms does, or than “street vendor” and the other two terms do. It’s possible that Google might set a certain threshold amount that candidate synonyms or substitute terms might co-occur with those other terms, and decide that “restaurant” is a much more likely and probably replacement than “store” or “street vendor.” So, Google may modify the query to search for “restaurant and chili and microbrewery beer” when the original query was “tell me about a place where I can get great chili and microbrewery beer.”

    In the four recently granted patents I mentioned, information about other words that might be within search results for specific queries, or within query sessions are the focus of these types of possible candidate synonyms or substitute terms. Google might also look at things such as click selections and dwell time on search results for pages within query results sets for these candidate terms. This type of information could be used to determine how satisfied searchers might be for candidate replacement query terms. Some of this data might be used to create databases where rules about synonyms or substitute words might be kept which could be used to make it easy and fast to find these replacement terms.

    So, if the query includes “place” and “chili”, the word “restaurant” might be a good candidate replacement term for “place.” If the query includes “place” and “pistol,” instead of restaurant, the word or phrase “gun shop” might be suggested as a synonym or substitute term.

    Of course things like schema and structured data might play a role in how frequently a page might be selected in search results, or how it might be presented there. But it’s not one of the main or key elements in the reformulation of a query term that appears to be at the heart of Hummingbird, which is aimed at better understanding the intent behind a query, especially a complex one, where the words used within the query don’t necessarily have to match the documents returned in response.

    1. A.J. Ghergich A.J. Ghergich said ...

      Wow Bill thanks for such a detailed response!

  6. Lauren Francis Lauren Francis said ...

    I heard about this vaguely, but didn’t know about it being named Hummingbird. Interesting stuff. I’ll have to see how this works out on my Tumblr platform, which sucks for SEO.

    1. Ah yes, Tumblr. Love it in theory but it’s so tricky for SEO! Maybe things will change in the future. It’s a neat platform in general :)

  7. eMarketing Wall eMarketing Wall said ...

    […] Hummingbird: Move Over Caffeine, Hello Sweet Nectar, ghergich.com […]

    1. A.J. Ghergich A.J. Ghergich said ...

      It sure does :) I will use the translator in chrome and check your article out as well later. Glad you enjoyed ours as well.

  8. Asher Elran Asher Elran said ...

    Marketers will have to write the content that is more engaging, vs the content that had the right keyword density. I think this update will make a lot of webmasters to re-write their content.

  9. Kate! What a helpful group of opinions you’ve put together for us here. What a time-saver and valuable education. Thanks a ton! -Ron.

  10. Sindy Sindy said ...

    For those who run sites properly with frequently generated content this update won’t play any role. For example, I have http://www.imcredo.com/ website with blog connected to it, and Hummingbird didn’t affect it.

Comments are closed.