Designing for the Google generation

I just spent a few days watching people use a search interface for a site (lets call it Site X for now). What stuck with me the most was how different the design challenge was if the user was of the “Google generation”. Everyone showed the Google carry over effect – expectancy of high relevance, fast searches. But for undergraduates who had grown up with Google, it was as if that was the only experience they could deal with. Anything else was too complex, too slow, too not Google.

Take Participant Y (Lets call her Susanne). Susanne is an undergraduate in a California University. She needs to find some information on Site X – she knows it houses some special types of information, not found anywhere else. She does not know much about this topic.

Her first impulse is not to start on Site X, but to start with Google. She goes to the Google toolbar (other undergraduates went to Yahoo or Wikipedia) and types in an exploratory query. She scans the results, focusing on finding an authoritative site which will tell her more. One site looks promising. On clicking the link, she finds its an ecommerce site and bounces back to the search results, and types in a new query. This time the first result gives her what she needs – more information about the topic, and some ideas about words to search on.

Armed with this knowledge and keywords, she (finally!) goes to Site X. She tries a long, specific query. And gets “0” results (unlike Google, site search engines do not have an infinite amount of information, and are not good at guessing what you are looking for. Very specific queries help only when you are using the language of the site.) She tries a few times with other, equally specific queries – each time getting 0 results. Finally, she decides to type in fewer terms, and gets some results. Results look promising, but its more items than she wants to scan (60 items is “too much”!). She adds another word in the search result, a word she saw in one of the result. Searching again and again takes time, and she complains about download speed pretty often. But in the end she is able to find the information she is looking for.

Susanne never uses advanced search. Instead her queries get more and more specific. As her queries get more specific, the chances of getting 0 results also gets higher, and her frustration with the site increases.

Later she talks about how she would be unlikely to use Site X again since she could probably find good enough results just through Google itself.

Welcome to designing for the Google generation. Highly skilled at query development. Fast, impatient, and in a strange way – extremely inflexible. Nothing but a Googlesque interface will do.

7 thoughts on “Designing for the Google generation

  1. It is interesting how an interface that performs a skilled task automatically (and for the most part, invisibly to the user) can effectively “deskill” regular users and leave new users unaware of the skills necessary to viably perform a task.
    You see some of the same effects in automation (humans left “out of the loop” on the system’s functions) and with dynamic system interfaces that do not allow users to grasp the functionality of the system.

    It almost seems as if the “Google generation” effect represents a lack of search skill (I’d include query development in this – that is, query formation as one aspect of search efficiency). As a more efficient search engine has taken the work of search away from the user, the search engine becomes more of a black box -user inputs text and *voila* – out comes a search result. The user has no clue what constraints led to their hit or miss on their search.

    This could be very good for a search engine (this has to breed reliance on their product), but doesn’t say much for the average user’s understanding of the mechanisms that propel information search.

  2. You could also fairly say it was the search interface on site X that lacked the skills it needed to run the query the user wanted.

    They can’t both be psychic, and, more importantly, if you want customers, your services better be. You have to know:
    1. What they will want, and
    2. How they need you to reveal it to them.

  3. Well, the search interface on the site definitely lacked skills.

    But I wonder how effective even a well done (but more complex) refinement / expansion interface would be. Really seemed to me that it there is strong design style that people have become used to. And they will use it even if it might be more adaptive to use it in another fashion.

    For example, another group of users (researchers) seemed more willing to adapt to the interface. They were also the ones who were used to non-Google interfaces for more complex searches.

    Lets take Flamenco (the search/browse interface I worked on at Berkeley) as an example: its an effective way to do searches if you are willing to give a more busy interface a try. But I don’t think the undergraduates who we watched would like it – some researchers we showed it to loved it.
    (Flamenco is at:

  4. I’ve spent the last 6 years tweaking the site search engine for a website with highly specialized content and an audience of non-professional researchers, and we’ve seen this in the post-google generation as well (bascially in anyone who was introduced to ‘search’ through google – which is almost everyone by academics).

    There are ways to make a simple search work for people with google-tendancies even in a content limited site. For instance requiring all terms in a search to be present works great for Google who have a zillion documents, but when you have 50,000 you need to be more flexible (if there are 66% of the terms present in a 3 word search, bring it back!). It requires a lot of testing, a lot of programing and a very good (and expensive ) relavance engine, but it’s easier to change the way the search works than change the way people work.

  5. I guess your point is to make site searches better. For end-users, they could still dig through the site with Google without using the search available on site (if they did find the site of choice that holds info in the first place via Google):

    “some string”

    I use this a lot, and it works all the time.

  6. Hi Rashmi,
    Another Google-effect I have seen is that people don’t bookmark the same way anymore (if at all). They just remember how they got there and do that same search again. This is especially true if the user found what they were looking for on the first page of the google result set. Admittedly I even find myself doing this too.

    I was watching my 8 year old niece do something like this.
    “Uncle Dave, I found this great new game.”

    “Oh! where did you find it.”

    “Let me show you” (she had no concept of the addressbar at all)
    – clicked google favorite
    – searched for “games”
    -knew exactly where on the page the link she wanted was (BTW, it was Yahoo Games)
    -went to Yahoo Games
    -found link to another site
    -went there directly to the exact game she wanted

    She was completely satisfied with doing this flow over and over again and when I suggested bookmarking the page she said why?

    Mind you she did have bookmarks, but only in the instances when someone else took her directly to a site (like her cousin, or her uncle) who typed in the URL directly. Otherwise, Google was her navigation tool, no matter how many links it takes.

  7. It’s interesting you mention the undergraduates and inflexibility. From my observations the undergraduates by far exceeded any other demographic in actually finding what they were looking for, often in the shortest amount of time as well. This was true whether they used Google to find it or using the site search.

    Many compensated for the poor site search by using Google to broaden their domain knowledge and attack the keywords on the site search with words or phrases they saw in the first several Google results. This worked fairly well, I’d say.

    The older demographics seemed to have more trouble completing some of the simple search tasks. Several failed to locate items they knew were in the database even with their extensive domain knowledge of the subjects the site covered. As a whole, they took much longer to complete the search tasks than the undergraduates.

    This is peculiar given the fact that many of these folks are doctoral level academics, people who spend much of their life searching in order to publish their research.

    Their lack of ability to adapt to the new Googlesque interfaces highlights their inflexibility. Add to that the fact that they took much longer to find items on the site on average, and I would conclude the undergraduates are definitely _better_ searchers.

    We’ll just have to wait and see how the Google generation has adapted to search technology in 2060.

Comments are closed.