Every ranking on WP Collections is built from data, not from who asked. This page explains exactly how, so you can check our work.
Where the data comes from
- WordPress.org. Active installs, ratings, number of ratings, last update, required WordPress and PHP versions. Pulled from the public WordPress.org API on the day we publish or update a list.
- Our lab. A fresh WordPress install on our own server, with no plugins and the same sample content for every product. We switch the theme or plugin on with its default settings and measure what a visitor’s browser downloads: the HTML and every stylesheet and script, gzipped.
How a score is worked out
Scores are out of 100. For themes the weights are:
- Satisfaction, 30. The WordPress.org rating, adjusted for sample size: we blend each rating with 25 imaginary votes at 80%, so a handful of perfect reviews cannot beat thousands of near-perfect ones.
- Page weight, 30. Home page HTML plus CSS and JavaScript, gzipped. 20 KB or less scores full marks; 180 KB or more scores zero; in between is a straight line.
- Adoption, 20. Active installs on a log scale: 10,000 is zero, 1,000,000 is full marks.
- Maintenance, 20. An update in the last 30 days scores full marks, sliding to zero at 365 days.
Plugin lists use the same idea with weights that suit plugins, stated in each article.
What we never do
- Take payment, free licences or anything else in exchange for a ranking or a better position.
- Rank a product we have a commercial relationship with. If we ever carry an ad for one, it sits in a slot labelled “Sponsored”, outside every ranking.
- Change a method for one product. If the method changes, it changes for the whole list, and the article says so.
Mistakes
Numbers go stale and we get things wrong. Email [email protected] and we will re-check and correct the article, with a note saying what changed.