A look at WordPress market share numbers

10 Comments

So just what is the WordPress market share? That sounds like a loaded question. Let’s break it down.

According to WordPress Stats, in September there were 27.7 million websites running on WordPress.com or using WordPress.org. The September 2010 Netcraft survey cites 227,225,642 hosted websites. So that means that WordPress makes up a little over 12% of all websites. Well, not exactly.

Netcraft also shows that of 227 million hosted sites, only about 96 million are active. Additionally, in March 2008, Matt Mullenweg was widely quoted for saying that about a third of all blogs were splogs, or spam blogs. He was really referring to WordPress blogs. The number came about because they had just “axed” 800,000 of 2.5 million WP.com blogs at the time. In his post he guessed that the actual number of splogs across the web was much higher, perhaps 80% or more.

In addition to splogs, I wonder how many WordPress installs are active? I’d venture to say WordPress has a much more active community than the competition, with a 40% market share (of blogs) and growing according to Technorati’s 2010 State of the Blogosphere. WordPress also has what looks to be about 50 times new blogs versus deletions. Even so, surely a significant percent of WordPress blogs go stale after a short run.

Sidenote: The NY Times did an interesting piece on why people stop blogging, and also quotes the 2008 Technorati State of the Blogosphere that said only 5% of all blogs were updated within the previous 120 days before the study.

If we use the non-scientific number Matt mentioned in his splogs post, we can guess that about a third of WordPress non-splogs are actively updated and visited. If we extrapolate these .com numbers to .org, that gives us a little more than 6 million active WordPress bloggers (27.7 × .66 splogs × .33 active = 6.03).  Now let’s take the 6.03 million WordPress bloggers compared to the 96 million active websites according to Netcraft and we get 6.28% of all active websites using WordPress.

It’s important to note that I believe this is quite conservative, given that WordPress is actively searching for splogs, and the “third” of “active” bloggers I used could just as easily be half. So if we said only 15% were splogs and half of WordPress bloggers are active, we get 12.26%. Also, a significant number of Netcraft’s “active” sites could also be spammy.

Okay, I wouldn’t have gone through with all the assumptions if I didn’t have something to compare it against. Builtwith Trends uses the Quantcast top million websites list to calculate the usage distribution of different technologies. According to Builtwith, WordPress was used by 7.4% of the top million websites on September 1st. Additionally, Builtwith shows that 5.91% of the top 100,000 and 4.05% of the top 10,000 websites also use WordPress.

I actually think the numbers are higher, because they show a footnote saying the statistics “are generated based on what can be found on the homepage of a website. Some technologies may be under-represented if they use custom implementations or do not appear on the landing page of a website.” So depending on what percentage of WordPress installations are not on the landing page, it could be significantly higher.

Plenty of stats to consider

After all of that, I think it’s safe to say it’s very hard to tell exactly what market share of the web WordPress holds. But I think we can reasonably assume that it is somewhere between 7.4 and 10 percent of the top million sites, and perhaps more than 12% of all active websites.

So what about looking in the mirror? Do you have any inactive .com blogs or .org installations?  I am responsible for five installations, and only three are consistently active.

10 thoughts on “A look at WordPress market share numbers

  1. I’m no expert on statistics so these are just my own thoughts. Regarding the State of the Blogosphere survey; surveying only 7200 people doesn’t seem like the best way to get an accurate sample. These are the more educated or active members of the community, not necessarily the most representative of the community as a whole. Then I’d also argue that that this excludes the people who use WordPress primarily as a content management platform instead of a blogging platform. Personally I don’t think that a wordpress.com site actually counts because I don’t believe that Netcraft tracks each of those separately. (I could be wrong, couldn’t find any answers either way) If they consider wordpress.com one site that could make a big difference in comparing those numbers. (And technically they’d have to if you want to consider wordpress.com as one of the most trafficked websites)

    • Doug, interesting thoughts! I’ll see if I can respond one at a time…

      1) I’d argue that 7,200 is pretty significant for a survey. Most surveys I see are 1000ish sample sizes. You can usually lock down some pretty tight confidence intervals with those numbers, and Technorati shows them on their intro(pg. 3 I think) to the survey. Also in the intro they show the populations where respondents come from. I think you do have a point though, that the average respondent may be more active than a non-respondent. But, what other option is there?

      2) The survey is mostly used in the article to figure WordPress activity levels, which I’m forced to extrapolate to cms WordPress applications within .org. I have a hunch they wouldn’t be too different, and if anything, cms applications would be more active, making the estimation more conservative.

      3) I believe Netcraft does count .com accounts, mostly due to the note they made in their October report about it. As an aside, I didn’t use the most recent reports on Netcraft and BuiltWith so that I could measure the same time frame across the board, hence September.

      Thanks for the input!

  2. This is a nice write up of what is truly a very interesting space.

    I am intrigued that as yet we have not seen an open real piece if research from WP.org looking at exactly what you have analysed.

    Also, as WP is more of a CMS these days than a ‘blogging tool’ I would love to know the commercial split.

    i.e. the number of sites which are commercial rather than Blog/adverting as WP has a huge development and design/web marketing community. It would be of interest the market share split between the proprietary systems and WordPress.

    Still 6M websites that are active with either blog content/commercial content is an impressive User base.

    It makes perfect sense that Microsoft and other larger tech players are now looking at the benefits of WP and the WP Community… yes we at Integrati marketing firmly believe it is a community!

    Great post thank you!

  3. The methodology Netcraft uses is fascinating. Netcraft’s counts actually only count “active” sites so while splogs will be counted, idle sites will get the axe in the counting metrics. Splogs, when found on WordPress.com, will be deactivated and NetCraft will detect this the following month.

    Another estimate by a creator of some-consider-competitor, Drupal is that WordPress occupies 8.5% of the entire internet. Not sure what his methodology is though.

    Another interesting sidenote is that the latest survey results feature WordPress.com prominently (with respect to the counting metrics used) due to the Microsoft Spaces migration.

    (Disclosure: I work at Automattic, but I hope I wasn’t too biased ;-))

  4. Pingback: Celebrate doubled signups with snow on WordPress.com | WPCandy

  5. Pingback: Making money with open source | Ben Werdmuller von Elgg

  6. Note that microsoft live is, well, dead, and wordpress.com is one of the migration choices. so that will add a few million more sites. When I went to Drupal con in 2010 Dries himself had gathered stats and showed that WordPress was way ahead of Drupal and joomla.

  7. Pingback: The Booming Wordpress Ecosystem | Mearo.co.uk

  8. Pingback: Why you need not delete the “Hello World” post that comes by default in WordPress Blogs?

Leave a Reply

Please note that WPCandy is a moderated community.

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>