Over the holidays, I came across an interesting blog post on HackerNews. The TL;DR is that the internet’s awesomeness stems from open standards, such as hyperlinks. The author points to an example of someone discovering the ‘App Defaults’ project organically by navigating various web links. It’s an enjoyable read on it’s own and I recommend checking it out. However, what really got me interested was my own discovery of the App Defaults project. The idea is simple: publish a blog post that lists your primary, or ‘default’, app for a number of different categories. The author linked a feed featuring over 300 bloggers who’ve participated in the exercise. Reviewing some of the posts, I noticed that Password Management was one of the categories many bloggers were listing and I thought I’d like to get that dataset.
Crawling the Data
I turned to ChatGPT to write a script that would crawl each of the pages listed in the RSS feed. Using OpenAI’s API, it parsed the HTML on the page and return a formatted list of the categories and apps. Finally, it wrote the results to a CSV file. In very little time, I had a spreadsheet of 300+ entries (sanitizing the data took some effort because of the different formats and spellings on different blogs).
The results
Of the 324 blogs I crawled, 284 listed a password manager. Among these, 149 use 1Password, 53 use Bitwarden, and 35 use Keychain, with the rest scattered across various other solutions.
| Password Manager | Count | Percentage |
|---|---|---|
| 1Password | 149 | 52.3% |
| Bitwarden | 53 | 18.6% |
| Keychain | 39 | 14.0% |
| Keepass(X/C) | 15 | 5.3% |
| Enpass | 6 | 2.1% |
| pass | 5 | 1.8% |
| Chrome | 4 | 1.4% |
| Strongbox | 3 | <1% |
| Vaultwarden | 2 | <1% |
| Lastpass | 2 | <1% |
| Firefox | 2 | <1% |
| Dashlane | 2 | <1% |
| Secrets | 1 | <1% |
| Proton Pass | 1 | <1% |
Being interested in developer users, I looked at the 186 respondents that included an IDE or code editor. Counting the most used IDEs amongst the 1Password users, we can see, to no one’s surprise, that VS Code is the most popular. What is surprising is that Nova, a feature-rich code editor for macOS, is second.
| IDE | Count | Percentage |
|---|---|---|
| VSCode | 42 | 59% |
| Nova | 15 | 21% |
| Sublime | 5 | 7% |
| Neovim | 5 | 7% |
| VSCodium | 4 | 6% |
Patterns by Password Manager
The last thing I did was count the top apps in each category, sorted by password manager, to get a profile of the user. The results aren’t exactly scientific, but some clear trends emerged:
- 1Password users are more likely to be Apple users who pay for premium applications, such as Fastmail, Fantastical, and Nova.
- Bitwarden users lean towards Google products (although Apple is still represented) and use free apps.
- Keychain users seem to stick with stock Apple apps whenever available.
Some Caveats
This dataset is small, and it hasn’t been seriously validated (I was on holiday, after all), but it offers an interesting overview nonetheless. More importantly, it highlights how we can use new tools and creative methods to sample data and uncover patterns. The speed at which I was able to go from simple idea to a useable dataset was incredible.
If you want to dig into the data yourself, you can browse the results here. Maybe you’ll find some cool tools or someone interesting to follow.