21 Files towards efficiency

Trust me, I am a computer scientist. And I can share great news with you: you will perform much faster if you keep up to 21 files in a folder, not more. Keeping up to 21 files in a folder will do the following:

• It will keep your folder structures flat. If you keep 21 files in a folder, use 10 subfolders in each folder, and you need to store 5.000 files (which is a realistic number according to the scientists Henderson and Srinivasan) you will find every file after 3-4 clicks. Because you need only 3 levels of folders. For example, we often find three levels of folders structured like this: /project/tps/proposal . You may end up putting less files into a folder, like "3 files" but that will result in a much deeper filestructure, where you will often go into dead ends (ah, the TPS report was not in "reports" but in "project/tps/reports"). If you put more files into a folder, you lose time finding the right thing inside a folder. Keep your file structures flat, it helps (I will blog more on file structures).
• It will speed you up. A group of great researchers observed 296 people clicking through their structures and found out that it takes you 2.2 seconds to click from folder to folder, and each file in a folder will distract you a bit from what you are looking for and slow you down by 0.1 seconds. Based on this mathematics, they found that the perfect number of files per folder is 21.
• Everyone else does it. The same researchers observed that the mean amount of items per folder is 22 and that 67% of all folders contained up to 21 items.

So - what should you do now?

Bookmark this page, remember that "21 is the magic number" and when you reorganize your files the next time, flatten your folder structure into less folders and split up folders that contain more than 21 files. People usually reorganize when they move to a new computer, or clean up each year, this would be a good moment.

The "rule of 21" is a textbook example of useful research. I love scientific PIM research, especially when it is properly done and thus returns trustable advice. At Gnowsis, I interview people to find out what they need. I did the same for my PhD (pdf) in the last years. While we build the personal productivity assistant Cluug, we embrace the scientific method - all features of Cluug have been tested and evaluated beforehand, you can be sure that it has proven to be a productivity tool for others already. If you want to learn more - tell me something about your approach. I will do some interviews and usability tests next week and I can pass on some of the other good advice we found.