Major update to our version stats for PHP, MySQL, and WordPress

Today, the stats reflected on changed, dramatically. I’d like to explain why.

First, so they’re documented here, the numbers:

WordPress versions

Version Before After
3.0 16.06% 2.27% −13.80%
3.1 1.21% 1.25% +0.04%
3.2 1.38% 1.35% −0.03%
3.3 4.11% 2.96% −1.15%
3.4 4.47% 3.24% −1.24%
3.5 14.60% 7.21% −7.38%
3.6 6.95% 3.72% −3.23%
3.7 4.34% 2.51% −1.83%
3.8 11.66% 8.61% −3.05%
3.9 13.20% 13.52% +0.31%
4.0 12.53% 17.40% +4.87%
4.1 9.48% 35.95% +26.47%

PHP versions

Version Before After
5.2 31.76% 16.60% −15.15%
5.3 38.56% 38.45% −0.11%
5.4 25.01% 37.18% +12.17%
5.5 4.29% 6.52% +2.22%
5.6 0.39% 1.26% +0.87%

MySQL versions

Version Before After
5.0 17.84% 9.31% −8.53%
5.1 25.24% 23.80% −1.45%
5.5 51.87% 59.35% +7.47%
5.6 4.99% 6.84% +1.85%
MariaDB 10.0 0.00% 0.68% +0.68%

What happened?

Dion Hulse (@dd32) has been working hard to modernize our stats collection and processing. This stats page, which is updated daily, has as of today been switched to this new data.

Thousands of new WordPress sites come online every day. Some others, though, stop pinging over time. The new data only reflects sites that have pinged within the last few months. There were also plenty of other inconsistencies in the data that we’ve been able to resolve, which has resulted in numbers we feel are more consistent and accurate.

There are three specific trends to note:

  • WordPress 4.1: More than 1/3 of sites are running the latest version, not less than 10% as previously determined. By excluding sites that are no longer online, you can imagine why this percentage would go up.
  • WordPress 3.0 finally looks more in line with what would be expected. This data has been an anomaly for years. (We’ve suggested before that this data was likely invalid — a byproduct of some spammers.)
  • PHP 5.2 is down to 16.6% of sites. We’re in far better shape for 5.2 than previously thought, though 5.3 hasn’t changed.

As Matt shared at WordCamp San Francisco in October, we’re engaging individual hosting companies to move sites to the latest versions of WordPress, with a secondary focus moving sites to PHP 5.4+. I also expanded on our reasoning and efforts during my php[world] keynote in November. One-sixth of all sites running PHP 5.2 is still many millions of sites. If we move the PHP minimum version too early, we risk stranding millions of installs on older versions of WordPress.

So, I wish to note that this does not change our calculations for keeping PHP 5.2 as the minimum for WordPress core — we had these numbers available to us when preparing our 2015 plans.

There are a lot of great things in this new data set. Hope you find it interesting!

Updated March 2: I shared some more info with WP Tavern here. In particular, I answered Is there anyway we can see PHP/MySQL versions broken down by what WordPress version they are running on?

We’re still working on ensuring the numbers are stable. They’re pretty predictable: older WP versions have more people on older PHP and MySQL versions. Newer WP versions have less.

PHP 5.2 is at about 16% for all installs right now. It’s at about 10% for installs running WordPress 4.1, but because 4.1 is such a large part of the pie (36%), it’s the WP version with the most PHP 5.2 installs.

Our goals remain the same: priority 1 is to update old WordPress installs, priority 2 is to update old PHP and MySQL. Only once the numbers drastically move as a result of our efforts would any minimum requirement change. We cannot risk abandoning so many users on older WordPress and PHP versions.