What Forty New Skills Do to Category Averages

Adding forty skills moved the directory average down by 0.08. The category-level effects are larger and more interesting.

The other categories — dev, data, design, writing, video, finance — are unchanged because the batch did not touch them.

Why marketing moved most

The batch added eighteen marketing skills with a mean of 6.09, including eight e-commerce files below 5.0. That is enough to pull a 141-skill category down by 0.31.

Documents moved because four new skills average 6.42, against a 27-skill category that was at 8.22. Productivity absorbed eighteen new skills averaging 7.58, but the category is large enough (191 skills) that the effect is only −0.08.

What the average does not mean

A category average is not a verdict on the category. It is a weighted average of our curation choices. If we import a thin repository, the average falls; if we import a strong one, it rises. The number measures the directory we built, not the domain.

That is why we publish scores per skill and let you sort. The marketing category still contains excellent skills; it now also contains honest low scores.

The curation question

Should we import a batch that lowers the average? We think yes, if the skills are real and the scores are honest. A directory that only contains high scores is a directory that has stopped looking.

The alternative — import only the top half of each repository — would raise the average and hide the distribution. We would rather show the 3.2 next to the 9.1 and let the reader decide.

This is the same position we took in The Long Tail Isn't Winner-Take-All: the distribution is the information.