Hitwise: Search Queries are Getting Longer
According to Hitwise, search queries on all the major search engines are starting to get longer and longer (PDF). While the average search query is still around two words long, queries that are longer than four words have become increasingly popular over the last twelve months.
Hitwise's latest data also confirms that Google's market share in the search business is continuing to grow at a steady clip (9% year-over-year). Year-over-year, all of Google's larger competitors lost ground, though at least between December and January, both Yahoo and Ask.com saw a very minor increase in their market share.
Longer Search Queries
Year-over-year, using one and two-word search engine queries became slightly less popular, while the number of three-word queries remained flat. Instead, a growing number of users are now opting to use longer queries. Overall, longer search queries have increased ten percent over the last year.
This is an interesting trend, and it could be interpreted in a variety of way. This could mean that a growing number of users is finding less value in the search results they get from relatively unspecific, short queries. It could also indicate that users are becoming more sophisticated in how they structure their queries when they are looking for very specific answers.
Do you have a theory why more users are turning to longer search queries? Feel free to let us know in the comments.
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
My guess is that it has to do with users becoming more sophisticated searchers. No longer content to "see what comes up" when searching on 1 or 2 terms, users want to get right to specific information fast and get on with their task.
I think,search engine companies(google,yahoo,and so on) are getting more information from The Web everyday.
So,enormous results come from searching one or two words.Users can't get right result.
As a personal blogger, I've been looking at crazy long search strings since the beginning. Peeps type in questions like their machine can just give them the answers. Which often, it does.
Autocomplete in the search box. Duh.
You can get to "curious case of benjamin button" with "cur" at Google or "curi" at Yahoo.
So what does this mean? Searches getting better at searching? Engines getting worse at sorting data? Probably both. As the data hungry spiders crawl more and more of today's web I am willingly to bet it is getting harder to produce good, relevant results.
At the same time searchers are required to learn to become better at searching.
On another note with universal search being used so heavily now in Google and other search engines, maybe searchers are just being forced to try longer queries to find some relevant text based results.
I think there are two potential reasons for the increase in words per query.
The first is that the early and late majority are becoming more comfortable with search. It wasn't that long ago that many would type in something like "books+used" or other very rudimentary searches. Lots of preconceived notions on what a computer could understand.
But after a while they left that type of Boolean search behind and started to treat search like the oracle on the mountain - you could ask it anything and it could point you in the right direction.
The second theory is that people are becoming frustrated with the search results and are reformulating queries to get better results. There's an interesting Yahoo! research paper on search query structure and reformulation.
So, the short queries return broad results while longer queries return more narrow results and likely provide more value and benefit.
Of course it could be a mix of the two (or neither!) as well.
I just blogged about this topic (using the Hitwise data and a Yahoo! research paper) and how it will impact SEO and PPC campaigns.
The Future of Search is Numbered
1. Growing user sophistication in formulating more precise queries, driven in part by:
2. SEO spam, and
3. Search engines implementing auto complete features, both as a general feature enhancement and to help users navigate around the SEO spam that populate the first page for short, popular "head" search terms
I guess, it can also be due to the fact that many search engines are now slowly developing support for contextual & NLP. Better algorithms have encouraged users to submit complex queries.
People know they can find what they want when they define well their search terms.
People will try to find information from different angles if the previous searches didn't result in what they were looking for.
Conclusion:
The more content your website has the more likely to be found.
This is an interesting trend. I think people are getting more and more sophisticated in their searches and more and more frustrated by the pile of general results they tend to receive. Let's face it: even searchers who are not quite sure which gadget they want to buy or what news item they want to know about are very specific about the type of item or information they are looking for.
A follow up question I had in my mind was this: Of the people typing in 6 or more words into queries, how many of them are asking a fully formed question? And in that subset, how many are getting decent answers to those questions?
I think this is a very interesting trend. My co-workers and I have been discussing the effects of long tail key terms and this just adds another piece to the pie. I think that with more and more info becoming available, people are having to refine their search to bring up specific, relevant results. I know that I have been using long phrases (including dates and other specifics) more in Google, especially if my first search doesn't turn up what I am looking for.
This is an interesting phenomenon, indicating search needs are getting more and more sophisticated---general search engines, however, are not adequate to serve these needs. Great news for vertical search engines!
In traffic analysis of our apartment search engine http://apartments.cazoodle.com, we do notice that the keyword queries leading to referrals to our site are typically quite long, many at-least 3 words.
Govind
The internet is growing quickly - there are too many websites with similar content out there to get good results for a generic one or two word search query. When the internet was smaller, it was easier to type in a shorter search query and skim through results to find a niche result; now everyone's talking about everything and results are growing, so it's necessary to be more specific so the relevant results pop up on top.
When I search, I go one of two routes: if I'm looking for a quick, basic answer, like "How old is Cher?" I'll type in a complete sentence and search engines like Google and Ask.com show the answer at the top of my results page. But if I'm looking for detailed information, I'll type in random keywords strung together without articles, and unless I'm looking for a basic Wikipedia article, I use four or more keywords together to narrow down results and make it more likely that the site I need will be in the top three.
I think Walter got this dead right. It's largely the effect of auto completes attached to the search bars in browsers like Firefox. In particular because average query lengths have not changed, only that the outlier sizes have increased.
However, there is almost certainly an accompanying increase in user sophistication where querying is involved. And if this is not currently the case, it soon will be.
The problem underlying this phenomenon, as seen by people like Cass Sunstein, is the potential for users to predetermine the information they find on the web. The vast plain of information on the web is scarred along paths of self-reinforcing attitudes (ie, inter-linking organization sites, if-you-don't-believe-me-click-here links, etc).
To the extent that information seekers employ rudimentary search criteria, their imprecise use will yeild myriad result types. For example, querying "nanotechnology" will likely provide a range of attitudes on the subject of nanotechnology. If, on the other hand, more adept search users begin their information queries in ways that reflect their pre-conceptions (eg. "nanotechnology+devastation+grey goo"), the results they find will be more one-sided.
Additionally, communication research indicates that attitude reinforcing information is more self-gratifying. Since this type of focused querying is inherently more satisfying to a reader, they may not be driven to venture off their chosen path to explore dissonant information on a subject.
Increasing user sophistication - prompted by declining quality of results for shorter searches as well as using/trusting search engines for more complex tasks. On Jinni (http://www.jinni.com) we see that users adjust surprisingly quickly to the possibilities of natural language search for movies and TV shows, including quite subtle parameters like mood.
However, there is almost certainly an accompanying increase in user sophistication where querying is involved. And if this is not currently the case, it soon will be.