What plugins and crime have in common: Skewed distributions

Two weeks ago, Joost published a blog post about innovation in WordPress, on which Joost and I worked together. In it, we shared a table (Table 1) showing how WordPress plugin usage is distributed. Spoiler alert: a few plugins dominate everything. As in, really dominate. Like, “Google-in-search” levels of domination. This distribution sparked a peculiar sense of déjà vu. I did the exact same analysis when I was doing my PhD. Only, the subject of my analysis was very different.

A table showing cumulative WordPress plugin installations by number of installs. The top 100 plugins account for nearly 80% of all installs, showing a highly skewed distribution.
Table 1: Distribution of WordPress plugins by number of installs

Flashback to a previous life

Before WordPress, before SEO, and even before I wrote thousands of blog posts, I was a researcher. A criminologist. I spent several years diving deep into Dutch administrative data to write a dissertation on intergenerational criminal behavior. Yes, that’s right: I studied how crime runs in families.

The core finding? Crime isn’t randomly distributed across the population. Far from it. A small number of families account for a large portion of all criminal convictions. Sounds familiar?

Take a look at Table 2.4 from my dissertation (page 43): in our dataset of 7,987 children from 3,590 fathers, just 106 fathers (about 3%) accounted for over 41% of all convictions. That’s skewed. Very skewed.

A table showing convictions of children grouped by their fathers' number of convictions. A small group of fathers accounts for a large share of total convictions, indicating a skewed distribution.
Distribution of children’s convictions by their fathers’ conviction history.

The plugin distribution explained

Table 1 in Joost’s post shows that:

  • The top 10 plugins account for over 50% of total plugin installs.
  • And the top 100 plugins account for around 80%.
  • There are 72,000+ plugins, most of which are… let’s say, less visible.

In contrast, in the world of crime, even the most criminal families don’t reach those levels of dominance. The criminal justice system is skewed, but the WordPress plugin ecosystem is supercharged skewed. 

So what?

There’s a bigger point here besides the statistical curiosity. Skewed distributions affect innovation, visibility, and resources. In criminal justice, it informs where we target interventions. It’s the people who commit multiple crimes that need intervention, because that’ll have the biggest effect on the crime rates in society.

In WordPress, the distribution tells us which plugins get the most attention in the WordPress repo. It tells us which plugins find their way to the audiences and which ones never make it past page 3 in any search of the plugin repo. Plugins with a lot of installs and good reviews are shown much more often in the plugin search results. Understanding that skewness, and being honest about it helps us build better systems. As a community, we should help new, innovative plugins to get attention in repositories (and outside of them). External solutions like Ploogins are part of that solution, but they are not a full alternative to improving WordPress’ plugin search. WordPress really needs that innovation to stay relevant as a CMS. 

Also, let’s face it: data is weirdly poetic. Sometimes the same statistical pattern explains why kids commit crimes and why Yoast SEO keeps getting installed on every new site.

Conclusion

Most things in life aren’t fair. Or equal. But some things are more unequal than others. So next time you’re scrolling through the plugin directory, remember: you’re not just browsing tools. You’re witnessing a crime of distribution.


Leave a Reply

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

Looking for our logo?

You're in the right spot!
Go to our logo & style page.