Google users not your average Internet users

by Eszter Hargittai on December 9, 2005

IDG News Service has an article with results from a study conducted by S.G. Cowen and Co. about search-engine use by socio-economic status and Internet experience of users. The findings suggest that Google users are more likely to be from higher income households and be veteran users than those turning to other services for search. Finally some data on this! I have had this hypothesis for several years, but had no data to test it. I am usually frustrated when people make generalizations about Web users based on data about Google users (worse yet, Google users referred to their Web sites through particular searches) and this is precisely why. I did not think Google users (not to mention ones performing particular searches on certain topics) are necessarily representative of the average Internet user. (The report says very little about the methodology of the study so it is hard to know the level of rigor concerning sampling and thus the generalizability of the findings.)

Interestingly, the survey found that 52 percent of users cite Google as their preferred search engine, Yahoo! comes in at 22 percent, MSN and AOL at nine percent each, and Ask Jeeves at five percent. These figures are not completely in line with data about search engine popularity by number of searches performed (from a few months ago). The Nielsen/NetRatings figures are somewhat different with over 10 percent of searches (by US home and work Web surfers) perfomed on other engines. According to the current study, only three percent list others as their preferred engine.

Of course, these two sets of numbers are not necessarily at odds with each other. The percentages reported in the current survey consider “search engine of choice”, while the Nielsen/NetRatings figures are about all searches. The SG Cowen & Co study findings may just mean that people who prefer one search engine over another still use several. I wonder if the present study had any questions about the use of different search engines. (A study I will be launching soon does ask about this. I would love to hear about other studies that may have explored that specific question.) A study called “How America Searches” published by iCrossing last summer found that while 77 percent of respondents use Google at some point during their online activities, only 13 percent use nothing but Google for their online searches.

So one question then is whether people will be more likely to switch to Google as they become veteran users. It is hard to say. For one thing, whatever led people to switch to Google a few years ago may not push people to switch to it now. Perhaps more importantly, Internet adoption is not a random activity and so those who have gone online more recently differ from early adopters (e.g. income, education) in all sorts of ways so simply becoming veteran regarding years of use won’t make them identical to the early adopters and thus more years online may not mean a switch to Google. It will be interesting to follow all this over the coming years.

On a different point regarding the IDG article: there is an unfortunate use of the term “Net-savvy” in its title. The author seems to equate Internet experience (measured as years of use not frequency of use) with Net-savvy. Research I have done shows that years of use is not a very good proxy for Net-savvy. In one study, I found that number of years using the Internet is a weak predictor of Web-use skill (measured as the actual ability to find different types of content online, quite relevant to the topic of search-engine use). Self-perceived skill is a better measure, but still not as strong as an index of items asking people their level of understanding concerning various Internet-related items. Perhaps it sounds more interesting to say “Google users wealthier, more Net-savvy”, but it’s a leap from the data available in this study (or at least the data that are discussed in the piece).



John Quiggin 12.09.05 at 5:35 am

Presumably the proportion of post-Google users, preferrring alternatives like Technorati or direct consultation of Wikipedia, is still to small to show up in studies like this.


Chris Karr 12.09.05 at 8:51 am

A suggestion for a change in the title of this post:

“Google users not your average US residents”

If Google commands the majority of search market share among Internet users, wouldn’t its users be more “average” than most other search engines among “users” on the basis of larger representation?

Since Google’s users are not demographically similar to the average US resident, the suggested title above makes more sense, since presumably American Internet users are a subset of American residents. I don’t think that it’s too surprising that the higher incomes are better represented in the pool of users (and better represented among the pool of people who have been online longest), so I would suspect that most search engines’ average users would be in the higher-earning demographics than the regular population.

Do we have any information on what the “average user” was in this study and how far Google and other search engines deviate from that average?


Eszter 12.09.05 at 9:48 am

Chris, the title is how I meant it. Perhaps it requires some clarification to say “Google users not your average Internet users in terms of their demographics”, but that becomes a bit tedious.

Saying what you proposed wouldn’t be very interesting given that Internet users, on average, are not your average Americans either and we’ve known that for over ten years. I guess depending on the particular search engine preference/use breakdown (whichever they used here to calculate their figures) it could be possible to find that users of one search engine actually do represent your average American. That would require that a search engine appeal especially to lower income users to offset the higher income of the average Internet user (compared to the income of the average American household).


Chris Karr 12.09.05 at 10:36 am

I guess depending on the particular search engine preference/use breakdown (whichever they used here to calculate their figures) it could be possible to find that users of one search engine actually do represent your average American.

I guess my question is that if you found one search engine whose users reflected the demographic traits of the average user, why is this interesting? Or in other words, let’s assume that “Ask Jeeves” users (whatever that may be) were closest to the average in terms of demographics. I don’t see what it says other than the user demographics of “Jeeves” coincidentally match the computed demographic average. You could title this post “Ask Jeeves users are your average Internet users”, but I don’t see how that would be useful more than as a bit of trivia if Jeeves users are only five percent of the search engine user base.

Maybe I’m missing something here.


Eszter 12.09.05 at 10:58 am

The question you pose was not my motivating question nor did it seem to be a question of interest to you based on your first comment. (I didn’t mention it in the post, it came up in response to your comment.)

The reason it is interesting to note that Google users are not average – the point of my entry and a point I mentioned in the post above – is because so many people draw conclusions about all Internet users based on the actions of Google users. Yet this seems incorrect given that Google users are not representative of all Internet users.


Chris Karr 12.09.05 at 11:29 am

I hadn’t seen many instances of people doing drawing the conclusion that Google users represent all users, so I probably read a bit more into your original post than intended.

Thanks for the clarification.

Regarding the question in your post above about new users switching to Google, I would bet that for pure keyword searches, we see a drop in Google’s new adoption rate. Working against it are smaller and more tightly focused search engines and more less-skilled people buying computers with other keyword search engines preconfigured (MSN, Yahoo, Netscape, etc.) as the default. On the other hand, quite a few alternative browsers are shipping with embedded Google search widgets, Google is developing searches beyond keyword matching, and they’re fielding a top-notch research and development team. There are so many variables that it’ll be neat to see how it all plays out in the end.


BroD 12.10.05 at 8:00 pm

I feel a bit like the person described in a Atrios post who realized that folks without medical insurance didn’t have access to standard medical care. I mean, they don’t Google?! What DO the simple folk do?

Comments on this entry are closed.