Which search engine appeals most to U.S. searchers? Which service is the most popular elsewhere in the world? It depends on who you ask, and how they're measuring.
A special report from the Search Engine Strategies conference, February 28-March 3, 2005, New York, NY.
Representatives from leading online measurement firms lent new insights to questions of size and scope, as answering behavioral questions of interest to online marketers at "The Search Landscape" panel.
While they agreed on the high-level question of who leads the search wars (Google), the extent of the Google Domination Factor (a term coined by search marketer Mikkel Svendsen) was in hot dispute. But that was far from the only statistical trend these analysts attempted to shed light on.
comScore Networks data, at least for the US market, has tended to portray the search wars as a fairly close contest. Banc of America Securities, in a research report dated April 22, cites comScore data giving Google a 36.3% share at the end of February, up from 34.7% in December. Yahoo is close behind at 31.1%; MSN clocks in with a respectable 16.3% share (source: Forbes.com, Google is "Undisputed Leader" in Search Queries, April 22, 2005).
The other agencies represented on the panel, Nielsen/NetRatings and Hitwise, saw it differently. Hitwise claims that Google searches represent 55.5% of all searches performed in the US market, with Yahoo at 30.8% and MSN well back at 6.6%. Those figures appear to make more intuitive sense to many marketers who report anecdotally on the traffic they're seeing in their server logs.
NetRatings' Ken Cassar had the search market share numbers at 47% for Google, 21% for Yahoo, and 13% for MSN. Cassar had the courtesy to also mention AOL Search, at 5%, badly lagging the "all others" category (14%).
Why the disparity between comScore and the others? There must be something buried deep in the methodology. On the surface, the key difference appears to be that Hitwise has a wider sample size than comScore's panel-based measurement method, though it too uses large samples. There are also lingering questions about the representativeness of panels. Some panelists' spyware blockers now even try to block the measurement firms' software (even though they've consented to installing it).
There is probably more to it than that, though. For now, comScore could be giving MSN and Yahoo credit for "searches" that do not take place through the conventional search box format, while failing to give Google due credit for unconventional searches and growing usage stats of its various "portal" features. Either that, or comScore's panels do not include enough of the "heavy searchers" which their own data indicate are the "key drivers" of search usage today.
comScore's data do confirm the Google Domination Factor globally. Panelist James Lamberti told the audience that Google has 61% market share in the UK, 67% in France, and a whopping 81% in Germany. Yahoo is well back in a distant second place in all of these markets, with MSN, Ask Jeeves, and AOL vying for third.
Hitwise's Bill Tancer emphasized that Google yields a significantly higher ratio of searches to visits than MSN. This probably means that there are fewer "one time only" novelty searches being performed at Google, and thus more serious searching is going on there. Arguably, it could also mean that some users are having trouble finding what they want. The aggregate data wouldn't give us specific enough user information to be able to tell.
In spite of some serious differences in the key figures, the panelists appeared to agree on a number of questions, and rather than exhibiting professional jealousy, tended to quote liberally from one another's available data to make various points. Clearly all recognize the limitations and constraints involved in attempting to take an accurate read of online user behavior.
The panel seemed to agree that Google users show a demographic skew towards male and affluent users, that Yahoo skews towards younger and less affluent users, and MSN towards an older female demographic. However, these differences have become slight. It has become inadvisable to make too many assumptions about the demographic characteristics of users of these top search properties, particularly Google.
Tancer offered a surprising nugget about the emerging category of desktop search: it is popular with older users. He made the audience laugh with an anecdote about his mother. Finding out that she and many of her friends were early adopters and frequent users of desktop search, he had to know why. "Because older people lose things, dear," she replied.
Some other issues covered by the panel included: adoption of shopping search and other vertical search tools as alternatives to general search; local search; the measurement of offline and latent purchases caused by search listings; and the strength of pricing for search ads.
The panelists noted that certain vertical search categories were holding their own, but that they also depended heavily on general search to maintain visitor levels. The implication is that without Google, Yahoo, and MSN referrals, many vertical search services will lose users to churn, to banner ads and broadcast TV ads inducing users to switch, etc. The fact that vertical players are either "inside or outside the primary traffic flows" is clearly an instance of portal power in action, and probably explains IAC's desperation move in acquiring a relatively weak #4 player in the search wars (Ask Jeeves) in an attempt to control some of the traffic flows to its vertical properties such as Expedia.
Properties like Froogle, Hotjobs, and Yahoo Shopping will no doubt continue to be strong players, but only because they're integrated with powerful portals that can remind users they exist. Obviously this bodes well for companies like Google and Yahoo. It does not bode so well for new players in vertical search. Global search brands have enormous clout and resources to outlast and outflank many vertical players, at least in primary categories like travel, shopping, jobs, news, local, and personals.
NetRatings' Cassar described the bubbling innovation in local search, including pay-per-call ads. But he noted that the opportunity is still a nascent one: only 3% of all searches are local searches, as against 97% for all other types of search. By the time you read this, that number will likely already be out of date given the strong recent performance of Yahoo Local and Google Local. comScore's Lamberti agreed with the 3% figure, but his data showed that a "less conservative" approach to measuring local searches already puts the figure at about 10%. Moreover, he noted, 60% of searchers had performed at least one local search in the past month. This appears to indicate that there is plenty of receptiveness in the marketplace and room for growth in local search, particularly if the local search product continues to improve.
Lamberti pointed to strong growth in usage of a toolbar to perform a search: 136% growth year-over-year to January 2005. Anyone who has ever found it convenient to type a search directly into their browser toolbar will understand why.
The panelists strongly agreed that paid search has not yet received the credit it deserves for brand impact and for its causal impact in offline and latent purchasing. Although the total number of searches will not increase as quickly as it has in the past, growth in search advertising should be robust at least for the next two years as prices on ads continue to rise based on a recognition that search listings are not only a superb direct marketing tool, but create similar latent impacts and brand lift benefits traditionally associated with broadcast and print advertising. In that respect, and in spite of the contemporary drive to measure everything that can be measured, one of the truisms of advertising remains current: "When I advertise, the phone rings. When I stop, it doesn't."
The data are often conflicting, but to close observers, paying attention to the winks and nods and voice inflections of panelists, clear trends can be identified. The Search Landscape panel has become an integral part of the Search Engine Strategies landscape.
Want to discuss or comment on this story? Join the The Search Landscape discussion in the Search Engine Watch forums.
NOTE: Article links often change. In case of a bad link, use the publication's search facility, which most have, and search for the headline.