Which Open-Source CMS Has the Most Active Development Community? [Part Two]

In Part One we estimated the number of websites that use Drupal as of July 2010. You can read that.

Per capita (so to speak), which of the big three open-source content management systems (Drupal, Joomla, and WordPress) has the most vigorous development community? This isn’t about the biggest community, but the busiest. For reference, here’s the estimated number of sites powered by each of the big three as of July 2010:

I’m going to keep this post shorter than the previous one, mainly because there’s much less estimating to philosophize about this time. Like our first post, there’s no direct way to quantify something like this, so we’ll have to fly at it from various angles.


One way to measure developer industry: the number of extensions each community currently offers. At the moment, here’s how many community-developed CMS extensions are listed by each community’s official repository:

Drupal: 6,190
Joomla: 5,274
WordPress: 10,278

WordPress has almost double the extensions of either Drupal or Joomla. But we’re not evaluating numbers without context; we’re interested in how many extensions each generates relative to its community size.

According to this, we can reasonably conclude the average Drupal developer is somewhat more likely to work on an extension than the average Joomla or WordPress developer. For the value of such an observation, I turn to you, comment section.

Of course, this isn’t a comment on the quality of these extensions. All three have pearls of really stunning work, and all three have duds and abandoned projects. It would take a million monkeys on a million keyboards a million years to learn how to operate a single module, let alone rank 21,742 extensions by quality. It would take a million people much less time, but there’s only one person writing this post, so any rankings based on quality won’t be considered here. Only sheer volume.


Did I take the time to add up the posts found at the official community forums of Drupal, Joomla, and WordPress? Afraid so. Totals:

Drupal: 836,000+
Joomla: 2,042,000+
WordPress: 470,000+

Totals in the context of community size:

So Drupal and Joomla people like to talk to other Drupal and Joomla people all day long, while WordPress people just blog it out. One could point out that Drupal and Joomla are much more complicated and powerful than WordPress, so of course more message board questions, strategies, theories, taboos, memes, and myths will arise from their respective camps.

Conference Attendance

All three communities have meetups around the world, whether they’re called WordCamps, Drupalcamps, Joomla! Days, or what have you. All three communities have these things several times a year in cities around the world. Attendance at these deals tends to range in the 100-300 ballpark. We’ll call attendance even for all three. Seriously, Google your mind out if you want, but it’s pretty close across the board.

However, Drupal also holds DrupalCon twice a year (once in North America, once elsewhere). WordPress and Joomla don’t have anything that compares — the most recent DrupalCon, an April weekender in San Francisco, was the tenth such event (we attended #8) and featured around 3,000 attendees, while the inaugural Joomla! World Conference has been postponed. In fact, DrupalCon has been such a success that WordPress is considering following in Drupal’s footsteps:

I’m not going to add up the claimed attendance of the dozens of Temples of Joom! and WordFests and Drupaloozas that have popped up all over the world, but DrupalCon makes it pretty clear Drupal developers are the most likely to gather with each other in large masses to be nerds all day. Until WordStock takes off, at least. And considering Drupal is the smallest community as far as user base goes, that’s pretty impressive.

In Conclusion

I feel we’ve made a non-insane case that Drupal’s development community is the busiest of the big three’s. While all three are amazing products and capable of doing just about anything, we’ve chosen to hitch 70% of our wagon to Drupal for a reason. (30% of our wagon remains hitched to WordPress, so here’s hoping they don’t go and call their conference WehatedrupalconCon.)


So that’s developers. What about designers? This one’s a blowout.

I Googled X themes (X being WordPress, Joomla, or Drupal) and noted the number of results for each — 775,000 for Drupal, 4 million for Joomla, and almost 54 million for WordPress. Divide those by the number of sites for each, and we have:

Even though wordpress themes is a popular spam thing, surely inflating those Google results, it’s still apparent that WordPress is the most designer-y CMS. Not that we didn’t know this, but here’s a little visualization of the nerdiest possible way to perceive the disparity.

Unveiling the Best Magazine Articles Ever and the Mood Swings of America: Three Best Things 7/26/10 - 8/1/1

Thing: The Best Magazine Articles Ever by Kevin Kelly. KK is putting together a list of the best magazine articles ever, spanning from an 1816 article on criticism to a piece from tomorrow’s New Yorker on hospices, and people are voting on them. This is the most obvious instant bookmark you’ve come across in quite some time.

Thing: Your Complete Guide to Bad Burqa Puns by Muslimah Media Watch. For newspaper headline editors, stupid burqa wordplay has replaced meet the new boss, same as the old boss as the cliche of choice.

Thing: Mood, twitter, and the new shape of america by Harvard’s Complexity and Social Networks Blog. Some math people did a Twitter-data thing to map our state-by-state zeitgeist as it changes throughout the day and across the country. You can read about how they did it, or you can watch the mesmerizing video:

Conclusions: people in Florida and California are pretty much never unhappy; Georgia is happy but the relatively grumpiest state in its neighborhood; and the Mississippi delta region and parts of the midwest are pretty much never happy.

America’s favorite time of day: quittin’ time is an obvious favorite, but early birds (people up between 5 AM and 7 AM) tend to be obnoxiously chipper* and have skewed our whole mornings green. America’s least favorite time of day: the post-lunch-pre-quittin’-time death march is pretty bad, but apparently oceans of horror start washing all over Twitter after 1 AM. Except in Florida and California, where they just have oceans of warm water.

* Full disclosure: I’m an early bird.

How Many Websites Use Drupal? Let's Estimate A Number [Part One]

We work with WordPress and Drupal. According to, there are 25.2 million WordPress sites as of July 2010. How many Drupal sites are there?

Very short answer:

Nobody knows for sure. But we think at least 7.19 million websites use Drupal as of July 2010, including hundreds of the world's most prestigious sites. Whatever the exact number is, it's growing exponentially.

Very long answer:

Since there are no exact numbers and no perfect ways to find exact numbers (in fact, the community lost count pretty much right away), we have to attack this from a few angles.

Angle 1: Compare Drupal's Market Share To WordPress' Market Share

Here's how the software used by the 10,000 most popular websites breaks down, according to Backend Battles, which is a website technology monitor and not a dance contest:

141 of the 10,000 most popular websites run on WordPress. 39 are based on Drupal (making them the web's two most popular content management systems). So Drupal has 27.6% of WordPress' presence on the Backend Battles list.

BuiltWith Technology Trends is similar to Backend Battles, but uses a larger sample size of 100,000 sites. According to BuiltWith as of July 2010, WordPress is used by 3.09% of websites, and Drupal is used by 1.67%. BuiltWith suggests Drupal has 54% of WordPress' user base. (Grain of salt: BuiltWith has been known to think everything's Drupal, like it's a kid in Sunday school answering every question by saying, "Um, the Bible?", but they claim they've fixed that by now.)

If we extrapolated Backend Battles' 27.6% and BuiltWith's 54% to the entire web, we could suggest that there are either 7 million or 13.6 million Drupal sites on the entire web. An average of the two would be 10.3 million. But that's likely too high. More on that in a second.

We'd arrive at these numbers by applying the Backend Battles and BuiltWith percentages (27.6% and 54%) to the number of total WordPress sites among the entire web (25.2 million as of July 2010), not just among the most popular 10,000. Because if Drupal has X% as many sites as WordPress among a sample size of 10,000 sites, it's reasonable to think the percentage would stay similar even among a set of all sites.

10.3 million: too high or too low?

On the one hand, the total number of WordPress sites includes both self-hosted sites that use WordPress software and blogs hosted on The sites measured by Backend Battles and BuiltWith are the web's most heavily trafficked sites -- almost all of them are self-hosted. Most blogs hosted on are going to be less popular than the web's 10,000 or even 100,000 most popular sites. The density of total WordPress sites is only going to swell the deeper we go past the most popular sites. Anyone can start a blog in seconds, and then never touch it again. Surely that's boosted WordPress' numbers quite a bit.

Drupal, however, doesn't have any widely adopted instant-website solution. If you want a Drupal website, you've got to really want a Drupal website -- committing to downloading, installing, and setting up Drupal is a bigger barrier to entry than WordPress usage faces. (Drupal's trying out Gardens as a quick-and-easy, WordPressy, hosted-for-you deal, but it's not likely to catch on because the freaking URL is

Imagine telling someone in a loud bar, "You should check out my blog. It's at"
Their reaction would be: "Partyface what gardens dot com?"
You: "D-R-U-P-A-L. I think it means rain in Norweigan; it's got modules."
Them: "Partyface dot D-R-U-P-A-L dot gardens dot com?"
You: "Close, but no. Let's walk through this one more time."


On the other hand, as this Joomla chart suggests, open-source CMS usage tends to be more prevalent once you venture past the top 100,000 sites, though perhaps none more so than blogs.

Brain hurts. Here's what's happening: cutting that Drupal average by 20%, a totally arbitrary number, to account for the thing.

Angle 1 result: 8.24 million Drupal sites.

Another web software monitoring site we considered, W3Techs, was way out there: Drupal has only 11% of WordPress' presence. On its face, that's not totally unreasonable -- 11% of WordPress' 25.2 million sites would suggest there are about 3 million Drupal sites, which is much lower than the others but not outrageous. However, they're only counting domains, not subdomains, meaning their numbers would only give WordPress credit for its 13.8 self-hosted sites, not its 11.4 sites. And 11% of 13.8 million is only a million and a half A MILLION AND HALF. As we'll get to in Angle 3, Drupal was downloaded more times than that during the third-biggest year of its shelf life, let alone its other eight and a half years A MILLION AND A HALF.

Angle 2: Look At What People Are Searching For

According to Google Trends as of July 2010, WordPress gets searched 3.2 times more often than Drupal does. Even though WordPress appears to be soaring while Drupal plateaus (we’d prefer the term consistent, even though Drupal clearly was Googled more in 2009 than it was in 2008), the difference is between 2.1 x and 3.5 x for every calendar year so far. Therefore, WordPress has had about 3.2 times the popular interest Drupal has had. Popular interest is relevant here because more searches are going to mean more downloads/signups YES I KNOW not directly, but I can’t imagine how there wouldn’t be some kind of correlation there. (Then again, Joomla was searched more often than WordPress was for three straight years, and there aren’t as many Joomla sites as there are WordPress sites; Joomla has been downloaded 14 million times since 2007 according to as of May 2010 in a long string of prepositional phrases without a comma. A fine number if it’s accurate, but not enough to approach 25.2 million. Sometimes all you can do is throw your hands in the air and proceed as planned. Watch me do that in the very next paragraph.)

Assuming X number of searches translates to Y number of new sites, we can again use the 25.2 million total WordPress sites as a base point. Dividing 25.2 million by 3.2 gives us 7.9 million, remarkably close to our 8.24 million estimate from Angle 1. Now let’s dock it by OH GOSH SAY 20%, because of the previously noted thing.

Angle 2 result: 6.32 million Drupal sites.

Angle 3: Expand On What We Know About Drupal Download Stats

Our first two angles depend pretty much entirely on the data that the WordPress organization has chosen to release to the public. They’re not known exaggerators, but our whole house of cards would be zoning violation’d if their numbers were off. So let’s try and do one independent of WordPress.

From 2006 to 2008, Drupal founder Dries Buytaert posted annual download stats, but hasn’t done so since. That’s fine. As with Joomla and WordPress above, we’ll assume they’re accurate.

Drupal downloads per year:

That’s almost 2.5 million downloads right there. More than two years have passed since the most recent update. If downloads continued at 2008’s pace, that would equal 5.7 million downloads, plus whatever happened between 2001 and 2005 — let’s say ~300,000 total for Drupal’s first four years. That might seem conservative, but it still adds up to 6 million. Of course, there’s reason to believe (based on our sources for Angles 1 and 2, plus other factors*) that demand for Drupal has risen in the past two years, so 6 million is too low.

* BuiltWith finds Drupal to be among the web’s 50 fastest-growing technologies between April and July 2010. DrupalCon attendance continues to double. 20% of Drupal’s total module projects were begun in the past eleven months (4,600 in August 2009 according to Wikipedia, 6,190 on as of July 2010). The many high-profile organizations that continue to convert to Drupal (in the past few weeks: the U.S. Department of Commerce, South Africa’s World Cup site, and Christina Aguilera, among others). And so forth.

Sure, a Drupal download doesn’t equal a new Drupal site. There are many reasons why a download wouldn’t end up becoming a full-fledged site — people could download just to try out Drupal or test a module, for example. But let’s say this is balanced by: (A) one download could spawn 1,000 sites if a developer was so inclined — none of which are counted here, (B) there are plenty of other places besides from which to download Drupal (if we felt like it, we could host it for download right here), (C) there are dozens of custom Drupal distributions like Acquia, iSite Essentials, Open Publish, and Pressflow that aren’t counted (D) alpha, beta, and release candidate downloads — most of which could yield a fully functioning site — aren’t counted, and so on.

Bumping that 6 million up by 1 million to account for Drupal’s increase in popularity since 2008 is very reasonable. Downloads almost tripled between 2007 and 2008, and we’re certainly estimating a much less aggressive rate of increase than that. Yep, that puts us right within range of our previous two estimates, which might seem like we’re forcing the numbers to fit. But, using 6 million as our base number, how many do you suggest we add to account for Drupal’s mini-surge of 2009 and, to a lesser extent, 2010?

Angle 3 result: At least 7 million Drupal sites.

Crappy, Discarded Angle: Search For How Many Sites Still Have The “Powered By Drupal” Badge

Every stock Drupal install includes a little footer badge with the Drupal logo and some alt text that reads, “Powered by Drupal, an open source content management system.” Most people take it down; some people leave it be. You’d think searching for this would be useful — no, not really. Depending on the search engine (and, seemingly, the time of day, ozone levels over Nepal, and LeBron James’ number of Twitter followers) it returns anywhere from 60 million results to 40,000. There are nowhere near 60 million Drupal sites that haven’t removed their showroom stickers yet, and there are a whole lot more than 40,000.

Even using Yahoo! Site Explorer and filtering to only include sites that link back to, as the badge does, isn’t helpful. I tried several different variations of the same kind of search, all of which wound up under 100,000.

This angle is like 89 degrees — it ain’t right. (And yes, this means there weren’t 40 million freaking Joomla sites two years after it was created.)

Angle 4: Fly Everything Up The Flagpole And See Who Salutes

Averaging the results from all three previous reasonable angles — because why not? — gives us 7.19 million. And until somebody comes up with a better idea, that’s our answer. Please come up with a better idea.

We estimate there are about 7.19 million Drupal sites, and we’re not as un-confident about that as you might think we should be. Your turn?

(7.19 million, and we just want to build you one.)

Three Best Things 6/21/10 - 6/27/10

THING: Sergey Brin’s Search for a Parkinson’s Cure from Wired. Google co-founder Sergey Brin, who’s made a life of using the power of algorithms to “organize all the world’s information,” is using that same kind of power in an attempt to cure Parkinson’s Disease. Since discovering he carries a gene mutation that puts him at risk of contracting the disease, Brin has sought “to bypass centuries of scientific epistemology in favor of a more Googley kind of science. He wants to collect data first, then hypothesize, and then find the patterns that lead to answers. And he has the money and the algorithms to do it.”

THING: Travel itineraries from Flickr photo trails from Geeking With Greg. Greg Linden links to a paper by American and Israeli researchers on “cleverly [using] the data often embedded in Flickr photos (e.g. timestamp, tags, sometimes GPS) to produce trails of where people have been in their travels.” It makes sense that the most interesting points along a path would also be the most photographed, so this could be a great way to note can’t-miss spots, common travel routes, and typical trip durations. An inspiring quote from the paper:

By aggregating such timed paths of many users, one can construct itineraries that reflect the “wisdom” of touring crowds. Each such itinerary is composed of a sequence of POIs, with recommended visit times and approximate transit times between them.

THING: How Rap Tears Up the Boring Art Vs. Commerce Argument from The Awl. Selling out has a different connotation in hip-hop than it does in other spheres. Though punk and hip-hop grew up at the same time and place and in the same socioeconomic conditions, the two have had very different ideas on mass appeal. (Yes, this is lumping thousands of musicians and millions of fans into two groups. I’m sorry.)

We can all agree the Black Eyed Peas sold out — they completely changed everything about their sound and image, conscious of their brand and marketability the whole time, and wedding receptions will never be the same. I’m sure they weep into their pallets of Franklins every night, thinking about all the underground respect they lost in the process. But many rappers have been able to market themselves without significantly changing their sound. None of this is new information, but the Awl article certainly presents a worth-reading take on the issue.

A Video of Americans Pretending to Care About the World Cup

Even though the U.S. lost its knockout round match against its nemesis Ghana, this World Cup still produced one of the best moments in American soccer history, and certainly the most widely experienced — Landon Donovan’s last-minute, life-or-death goal against Algeria did the kind of Twitter damage unseen since Michael Jackson’s death. An amazing montage of Americans from Arkansas to France celebrating the score:

World Cup Fun With Language: How Many People Call It 'Soccer' Instead of 'Football?'

It may seem Americans are the world’s lone gunmen when it comes to using the word soccer instead of football, but Canda, Australia, and New Zealand are guilty of soccrilege (anybody?) as well. In fact, at least one-fifth of the English-speaking world calls it soccer, not football:

All data from Wikipedia. I broke the countries down as follows:

Football countries: England, Wales, Scotland, India, Pakistan, Sudan, Nigeria, Cameroon, Uganda, Kenya, Botswana, Zimbabwe, Liberia, Sierra Leone, Guyana, Madagascar, Trinidad and Tobago, Jamaica, Malta, Singapore, and Belize.

Soccer countries: Ireland, New Zealand, Australia, Canada, the United States, and South Africa.

I tried to avoid English-speaking countries where there’s no clear winner — for example, the Philippines. There may be some debate over including South Africa as a soccer country, as its national team calls itself the South African national football team. However, the domestic league is called the Premier Soccer League, and the 2010 World Cup flagship stadium is called Soccer City.

Any disparity introduced by including South Africa as a soccer country is more than outweighed by this: not everyone in India speaks English as a first language — many Indians don’t speak any English. The same goes for many of the African countries listed as football countries. So if we were to get particular and exclude South Africa, we’d also need to ding India’s football population by a significant percent.

An 80/20 split means everybody should be happy. Football proponents can rest easy, knowing the beautiful game’s traditional name is securely dominant. And there’s comfort for users of soccer in knowing we’re more than just a handful of stubborn Americans.

Though it’s often confusing and inefficient, there’s no real harm done by using two different names for the same sport. It’s just the way language works. Sure, the game hails from England, but Americans invented trucks, and you don’t see us getting mad when English call a truck a lorry.

The obvious, as John Cleese states, is that the game Americans call football involves at least as much hand as it does foot, and its principal object isn’t exactly shaped like a ball. (As Americans, we’d eventually retort by pointing out cricket doesn’t involve insects. But first we’d run a monster truck lorry into a wall of fried chicken, because so what?)

Some Extra Charts

For no good reason at all, let’s also look at the same two groups, using combined GDP as our metric:

And finally, an even less relevant chart: