Vertical Creep into Organic Search Results
Google doesn’t just have a search engine. It has multiple search engines. If you look at the tabs on top of the search box, you’ll see:
- Web
- Images
- Groups
- News
- Froogle
- Local
Each of those tabs generates a different type of search, with its own unique set of algorithms and ways of presenting the material. These are narrower searches and are often referred to as “Vertical Searches.” They are said to appear in SERPs through a process referred to as vertical creep.
There are many examples of limited-scope search engines like this on the web for things like jobs, travel, hotels, shopping, and much more.
Google doesn’t limit itself to those vertical searches listed on the tabs above the search box but has even more available. Here are some of the other areas covered. These can be found through the default web search box on the Google page.
Weather (type in – weather and a city name – weather Cincinnati)
Definitions (type in – define:xxxxx to see a definition result, type in this format – what is zoology for a vertical search result above regular results)
Music (type in a band name, for example – Foo Fighters)
Books (type in the name of a book, for example – Moby Dick)
Vertical Creep into Regular Results
Interestingly, sometimes Google will show results from the other tabs and the other vertical searches within their web search results. So being first in organic results means being first after these vertical search results.
These additional results can be news stories, images, sponsored ads, definitions, Froogle products, and others.
For instance, on a regular search for “Ken Griffey, Jr., I see an icon of a newspaper and the following news articles:
News results for Ken Griffey, Jr. – View today’s top stories
Griffey sparks rout by US – Miami Herald – Mar 11, 2006
Team Reaches Next Round Behind Clemens, Griffey Jr. – Los Angeles Times – Mar 10, 2006
Griffey stars as US dominates South Africa – Sports Network – Mar 10, 2006
Many folks in the search marketing industry have referred to this integration of vertical search results into organic search results as a vertical creep. The name probably came from the concept of “scope creep,” which expands requirements and objectives during a project beyond what was originally planned or envisioned at the start.
Google isn’t the only search engine to include additional search results with organic results like this. The practice started years ago with search engines like Altavista, including some news results at the top of search results pages. This movement of other search results into organic results has come to be known as a Vertical Creep.
What triggers this vertical creep?
Good question. At the recent New York SES, I attended a panel on Vertical Creep Into Regular Search Results, and the presenters gave excellent overviews into the subject. Still, they didn’t discuss how these results are initiated. A snippet from the recap at Search Engine Round Table:
So what is a vertical creep, and why should marketers care about it? Its been called invisible tabs, Google Onebox results, Yahoo shortcuts, AOL snapshots, and Ask Jeeves Smart Search. Even if users don’t choose to do a vertical search, there’s a good chance that vertical listings will appear at the top of regular search engine results. So, for example, do a search for “Danica Patrick,” and image results creep to the top of web results.
There is a Google patent application that describes this process of integrating vertical results into the regular results. It was filed with the US patent office on December 31, 2003, and published on July 28, 2005. The listed inventors are Bret Taylor, Marissa Mayer, and Orkut Buyukkokten. US Patent Application 20050165744 is:
Interface for a universal search
The abstract for the document gives us a quick glimpse into what it covers:
A search engine may perform a search for a user search query over many possible search categories. For example, the search query may be performed for general web documents, images, and news documents. The search engine ranks categories based on the search query and/or the documents returned for each category and presents the search results to the user by category. Higher ranking categories may be presented more prominently than lower-ranking categories.
Universal Search and Vertical Creep Helps Match a Person’s Intent While They Search
One of the main reasons for this vertical creep into organic results is that while some vertical categories may come closer to matching a person’s intent searching, searchers tend to ignore the tabs above the search box that will allow them to access these more specialized results.
The search engine reviews the different categories, or genres, of results after a query is made to see if they are relevant to a searcher’s query. The results from those are ranked and compared. Here’s a snippet from the patent application:
Ranking component 402 may base its ranking on the documents corresponding to the links in each list and/or the user search query. For example, ranking component 402 may generally compare the search query to the contents of the documents in each list and base its ranking values on the closeness of the comparison. Consider the search query “buy athletic shoes.” For this search query, ranking component 402 may determine that the user is most likely interested in selling athletic shoes. Accordingly, a ranking component may rank the “products” category highly. The links in the list that correspond to the products category are likely to be links that correspond to web pages offering shoes for sale.
In some instances, the inclusion of a specific word may trigger the use of a certain vertical search genre. The weather and dictionary (what is …) examples above, for instance. Or the use of a word like “buy” may indicate that the person searching is trying to compare products for sale.
So it appears that what triggers the use of vertical search results is a guess at the intent of a searcher based upon both the words that person uses and a determination that a certain category might be a good match for the search. For example, a search for lion shows us images of lions at the top of the results because the search engine decides that the images category is the best match for the intent of the person performing that particular query.
Being placed in vertical results
One of the first questions asked at the New York Seach Engines Strategies Conference during the question and answer session of the presentation I mentioned above was “How does a site appear in those results?” That was followed by “How does my site appear in those results?” The answer to the first question was that each type of vertical search has its algorithms. There wasn’t much of a discussion of those individual algorithms, though.
That’s probably a good topic for research and experimentation. For instance, a previous post, Looking at Google Definitions, provides some idea of how to get into definition results from Google. There’s been a fair amount of research conducted by many involving showing up in local results. There’s more about news articles in this patent application from Google – Systems and methods for improving the ranking of news articles. Why are some images presented and not others?
The other major search engines are also returning these types of vertical searches. The chances are that personalization of search will also influence these types of results in the future if they haven’t started already. It’s a topic that’s worth more than just one blog post.
We have referred to this as vertical creep, but it has developed into Universal search at Google.
Nice article Bill, those are always questions that we ask ourselves but never really bring to discussion. I agree that personalized search will greatly influence those types of results: we are already doing it with paid search (see AdCenter and now Google with demographics targets), so it should be coming soon with natural search.
Thanks, Nadir.
This is an area that people have been quiet about.
With the personalization of these vertical items, I can see better result and category matches if the search engines know more about you. When I type in “weather newark” I’m given weather conditions in New Jersey, but I live in Newark, Delaware. If the search engine knew that, it might have responded with the Delaware weather.
The same with matching up these results to topics I commonly search for.
Just how much information do we want to share with the search engines, though?
Like you said with the demographics in paid search, they seem to know more than a little about us anyway.
Check out my last post, Google also helps you book a flight, isn’t it a nice function?
Nice post. Interesting the different results you received from the .com and the .fr versions of Google, too.
Yeah, you know what, I know where it might be coming from: more people are using the train to go to London from Paris rather than the plane. The question is: Did Google know that and added this function “by hand” or did they realize that after gathering personalized information?
A third option? Mining the data they have, and analyzing it on a larger scale? I don’t know how much personalized information you need to decide between showing information about trains or about planes.