Hacker News new | past | comments | ask | show | jobs | submit login
What is Google's real market share in the US? (gabrielweinberg.com)
87 points by pardo on Aug 13, 2011 | hide | past | favorite | 34 comments



I used to work on search quality at Google, but I think I can answer this just using public information. ;-)

I think the answer is that Yahoo and Bing have a much higher rate of navigational queries - queries like [facebook] or [twitter.com] that are just the name of a site. This class of queries does indeed send a much higher proportion of their traffic to a small number of sites.

The available public statistics seem to confirm this. One example is here: http://blog.alessiosignorini.com/2010/02/average-query-lengt...

I'm sure they're not perfect, but let's assume these numbers are reasonable. Single-word queries are a decent proxy for navigational queries. 27% of Google's searches are one word, but 54% of Yahoo's searches are one word. So if you were measuring on any site that didn't get one-word traffic, you would see Google's searches overrepresented by 1.5X already.

This makes sense if you think about it. The sort of person that searches on non-Google search engines is very different than the sort of person that searches on Google. People who search on non-Google are much more likely to generally not understand the internet. They are likely to just use the search engine that was preinstalled when they got the computer. They are likely to not understand the difference between a URL bar and a search box. And they are likely to only use search for typing in the name of a site and directly going there.


> I think the answer is that Yahoo and Bing have a much higher rate of navigational queries - queries like [facebook] or [twitter.com] that are just the name of a site. This class of queries does indeed send a much higher proportion of their traffic to a small number of sites.

> This makes sense if you think about it. The sort of person that searches on non-Google search engines is very different than the sort of person that searches on Google. People who search on non-Google are much more likely to generally not understand the internet. They are likely to just use the search engine that was preinstalled when they got the computer. They are likely to not understand the difference between a URL bar and a search box. And they are likely to only use search for typing in the name of a site and directly going there.

From observing non-tech savvy people using the internet, what I've found is that a lot of times, they know about Google and use it when they want to consciously make a search. But when they just want to access a site, they will type just the name of the site into the URL bar, which with modern browsers increasingly results in a search.

This method of searching uses the default search engine, which is often altered from Google even in browsers like Chrome due to these non-tech savvy people installing various software that changes the default search engine. But when these people want to do an "actual" search, they literally go to http://www.google.com/ and type in the search query there.

What's interesting is that a lot of times, they'll accidentally search with the default browser search instead, and they don't even realize that they're not using Google, which leads me to believe that they don't realize that there are different search engines, or that depending on whether they use their browser's default search or go to http://www.google.com/, they will get different search results.

I would be really interested to understand what goes through the minds of these people when they use the internet, because these sorts of things just scream out at me when I watch them browsing the web, but they seem to be completely oblivious to it. I just don't understand how they can miss the completely different site layout and URL.


What goes through their minds? Probably something like "I want to see what my friends are up to" or "I want to post a picture I just took" or "I'm bored and want to play Farmville". They just want to accomplish their task and browsers/URLs/search engines/keyboards etc. are just a means to that end.

I could imagine a similar analogy would be someone who is a professional racecar driver or just really into cars would be driving on a road thinking "I'm going to ease into this turn, and stay close to the inside" whereas most other people would be thinking "I have to remember to buy milk at the store after I pick up Jimmy from daycare".


What is going through your mind when you type "gm" and then hit autocomplete? It's a memorized task.

Do you wonder how DNS resolves? Do you wonder how the certificates are validated? Do you know what's in your .pem file? Who cares? It works, right?

Maybe you could make it more efficient, with some launcher program. But tasks that are memorized are easy to do (from a cognitive perspective), so you don't bother trying to improve them.

The human brain can be the laziest organ in the body. Well, it can be hard working, but in many cases it gets crazy blind spots, and gets stuck at a locally optimum point.


> I could imagine a similar analogy would be someone who is a professional racecar driver or just really into cars would be driving on a road thinking "I'm going to ease into this turn, and stay close to the inside" whereas most other people would be thinking "I have to remember to buy milk at the store after I pick up Jimmy from daycare".

I think a more accurate analogy would be taking a different route to work and not even realizing it.


Then, for these users you describe, it could be that every google search is preceded by a bing search (from the browser search bar, for "google").


I used to work in Search at Yahoo! and can say that you are correct.


This is an excellent point, but I think you mean that when taking non-1 words into account it would show that Google is about 1.5X underrepresented. If indeed that is the case, then the true "search/discovery" market share is about 90%, as supported through other commenters' site data.

Also compelling is the data AJ007 presents on number of days/month used. Bing at 8.9 trailing Google(21.5) and Yahoo (18) significantly. To me this clearly suggests that Bing users are the relatively unsavvy searcher of the Bing/Yahoo bunch. Does anyone have any recent demo data on the search engines? My guess is that Bing has a relatively older mix.


I think the answer lies in several areas, yours probably explains 90% of it.

Other things:

-Comparison of average number of visits a month per unique user: Google-21.5, Yahoo-18, Bing-8.9

-From what I have seen with my companies own internal numbers, Yahoo & Bing users may be more inclined to click on search ads. Unless a webmaster is running a big ad campaign, they aren't going to see huge volume from these two search engines.

-What percentage of Yahoo's 140m monthly users are visiting the site to use Yahoo Mail verse to search?

I recommend using compete.com for traffic intelligence. It fairly accurately reflects my own site's data. I've been paying $500 a month for their lower premium tier for a good 2 years with no complaints for anything other than idiotic salespeople.

Another comment, as a webmaster comparing your percentage of traffic from one search engine verse the other is meaningless unless your rankings are exactly the same in all of them. A real test would have to be done with a PPC campaign strictly receiving traffic from google.com, bing.com, and yahoo.com


I'm in sales and it just cracks me up when I see sales people who are selling a technical product do the stuff you pointed out. It is just a crackup and I try to show them there are better and faster ways to do things...Like opening multiple tabs vs. opening 4 or 8 browsers... And if you show them they forget. It is amazing.....


The article mentions "What makes this even weirder is Hitwise also recently came out with a report saying Bing/Yahoo users click on more links than Google users." This is a metric that Hitwise calls success rate and they say Yahoo has the most successful searches.

I did a debunk of Hitwise's "success rate" metric back in February: https://plus.google.com/109412257237874861202/posts/A7d5joZ3...

In essence, I don't think Hitwise can tell a the difference between a click on a search result vs. just going to a different site. So doing a search and then leaving looks just like a success to them. For example, yahoo.com has a "Trending" box in the top-right of the page, and clicking on Rose McGowan will do a Yahoo search. If the user goes anywhere after that search, I believe Hitwise will call that a successful search. Likewise on msn.com, if you click many of the links, e.g. "Pencil skirts that will turn heads" that actually does a search on Bing. I believe going anywhere else on the web afterwards will again look like a successful search to Hitwise. I think a large amount of casual browsing on Yahoo/MSN is counted as "successful searches" by Hitwise.


My programming blog gets 99.35% of its search engine traffic from Google.

AnyNewBooks.com: 96.44%

My math blog: 92.91%

OK, those are technical or for a tech savvy crowd. How about my wife's vintage blog which attracts mainly a non-technical female demographic? 84.08%. Or her Girl Guides' blog? 78.74%.

My father in law's local renovation site: 81.48%.

Not a single site out of a couple of dozen I've access to, shows anything less than 75%.

With these numbers at hand, I have a really hard time believing that Google's market share is anything below 75-80%.


Not many people realize this, but wikipedia makes all their stat information public. http://stats.wikimedia.org/wikimedia/squids/SquidReportOrigi... As a top 10 site with a pretty broad appeal I'd say they are one of the best gauges of internet users.

Google represents at least 90% of search engine references coming in to that site.


This is basically a measure of which search engine promotes Wikipedia links more heavily in their algorithm, not a measure of the search engine's market share....


If I'm looking for an article on something and Google thinks Wikipedia is one of the solutions it displays it amongst OTHER results. I click on it because I tend to like their articles.

I say the stat is what the stat it says it is. Wikipedia visitors tend to use Google more than other search engines to get there. I personally Google something with "wiki" thrown in as a keyword at the end so I can get to the correct Wikipedia article.


Is that a global number or just the US? If global, is it possible to segment to US?


I believe it's global, but if someone wanted to do the legwork to determine US only, all the log files are uploaded on an hourly basis here http://dammit.lt/wikistats/


I saw traffic numbers $HUGE_ISP last year and google was pulling almost 85% in search. That was by number of bits shifted, not impressions - I'm not sure how that would change it.

If you're really interested, I'd reach out to sandvine. They are on net at several majors and they gather tons of information and are allowed to share it. They might take some time out from writing breathless press releases about netflix traffic to kick you some numbers, all good publicity. I think volume on a large national ISP is likely to be a much better metric than trackers.


Could it be the result of Google lobbying for underestimating their search market share (due to antitrust concerns) and Yahoo/Bing are lobbying for overestimating their market share?


I don't think so (but your thought is very interesting anyway) because when Bing lobbies to get higher stats, they may get higher ads revenue but in the long run, particularly in the FTC investigation, they will run into bigger trouble.

But yeah, this is just my guess...


> That seems high indeed, but everyone I talk to "in the wild" who runs high traffic sites actually sees a much higher percentage of their search engine traffic coming from Google, usually from 80-90%.

Wouldn't the people that Gabriel Weinberg talks to be people who run websites much more likely to be frequented by the highly tech literate? Seems like a fantastically biased data set.


Nope, the people I was referring to for the most part run sites targeted at the general public not techies.


Yup. See my comment about my wife's blogs above. Sites for absolutely non-technical audiences get 80+% of all their organic traffic, from Google.


I have no idea who you talk to, but I'll say this: just because you target your website at the general public doesn't mean that your user bases isn't significantly slanted toward techies. It's not just the subject of a website that can bias is like this; it can just be their size, or their newness.


You can generally tell by browser share breakdown as I mentioned in the post wrt to statcounter.


Would like to see the equivalent of Japan ... Yahoo.jp has a ridiculous market share here (especially amongst the pre-smartphone market - which is huge).

Edit: best I could find was from 2009 which puts yahoo at 51% over google at 39%[1]. However, google seems to be partnering with yahoojp to help them improve their search technology[2].

[1]http://www.comscore.com/Press_Events/Press_Releases/2009/3/J...

[2]http://www.asahi.com/english/TKY201012020342.html


well on most SEO conferences there is the agreement that a good search optimized site gets at least 90 % from its search traffic via google in the us market.

in other western markets (i.e. germany) google has a market share of 98% plus and we are fine with it.

said that, one of my sites only gets 28% of its search traffic in the us market via google, but it has a penalty on most of its pages (everything but the startpage and its blog is banned from google and with good reason) so its an outlier.


as far as anyone is concerned Google does indeed have a monopoly on search...Bing spent something like 150 million to bump their share by a few percentage points

And it's not like Google's product is that much better than Bing...the difference is really miniscule...but that's the problem...Google is good enough so to beat them you really need to bring something truly revolutionary to the table


Is it just me or is Hitwise acting pretty aggressively/misleading against Google lately? It started last month with some of their Google+ stat numbers, and it's been at least 2 more such cases lately.


IMO, there is little question that Google's search numbers are higher. This is one reason why I think US antitrust probes are more justified given some of their actions (not saying they've done anything wrong necessarily, but prudence is advised).

This also explains why MS search revenue is lower than expected. People don't click ads in navigational searches.


I'll share some figures for some gaming sites from google analytics, these are global though, does anyone know how to narrow down by country with ga?

site a: 110m page views past 30 days, 51% Search engine traffic, google is 95%+

site b: 90m page views past 30 days, 77% Search engine traffic, google is 95%+


According to Google Analytics, Google is 97.3% of my search engine traffic.


Tech heavy site?

I'm doing 98.3% Google here. (ignoring the mysterious search engine called 'search')


Traffic sources for the last calendar month as measured by GA: [google, yahoo, bing]

my blog about R, bash, and whatever crosses my mind: http://blog.earlh.com : [1.4k, 9, 9]

my gf's blog about her life and our dog: http://danceswithcorgis.com : [877, 37, 14]

edit: see also acangiano's blogs: http://news.ycombinator.com/item?id=2881340




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: