Bundler.io Content Audit
Project components: User Research, Competitive Analysis, Content Audit, SEO Copywriting
Tools: Google Forms, Google Sheets, Google Analytics, Screaming Frog SEO Spider Tool
Bundler is the open-source dependency manager for the Ruby programming language. Its main site, Bundler.io, is also home to all of the project's documentation, including docs for older release versions. In total, the docs site contained over 200 pages of documentation.
The goal of this project was to create an inventory of existing documentation, understand which Bundler versions users relied on the most, identify the pages with the highest traffic, and improve docs' pages SEO.
Bundler's documentation landing page before the audit. (December 2017)
The project was divided into four phases: user research, competitive analysis, SEO audit, and content recommendations.
In phase 1, I developed a user survey to better understand how developers use Bundler in their projects, what tools they used to learn Bundler, and what resources they use when they need more information about Bundler. In particular, I wanted to get a better sense of how developers in the "learner" (defined as developers with less than one year of Ruby experience) and "proficient" (developers with 1-2 years of Ruby experience) categories used Bundler and what their pain points were. We found participants primarily through Twitter, and received 65 responses.
In our survey, I asked respondents to identify projects they’ve used that have useful documentation. Based on their responses, I focused the competitive analysis to four open-source projects: Django, Rails Guides, RSpec, and Vue.js. To collect metadata, I ran these sites through a site crawler (Screaming Frog SEO). I also visited their homepages to understand how their documentation was structured, how content was organized, and see key differentiators in their documentation.
Notes compiled from different open-source projects during the competitive analysis phase.
From the survey, we learned that:
92% of respondents use Bundler mainly in Rails apps, followed by 59% who’ve used it to create a gem. Rounding out the list were single-file Ruby scripts (56%)
and Sinatra apps (51%).
69% of respondents learned Bundler using the Bundler docs, followed by 25.8% who relied on the `bundle help` command to access information.
69.7% of respondents used v1.16, followed by 47% who used v1.15.
80.3% used Bundler in work (66.7%) and in consulting projects (13.6%).
66.7% of respondents turn to Stack Overflow to troubleshoot Bundler-related issues, followed by 50% who use Bundler’s docs for troubleshooting issues.
45.5% of respondents typically have to search for solutions to Bundler-related problems just a handful of times per year.
From the competitive analysis, we learned that:
All of the projects used metadata consistently. They employed standardized titles, included detailed meta descriptions, and utilized keywords.
These projects made it simple for users to jump from one version's docs to another.
Some projects encouraged contributions by adding calls to action on each page, which would take visitors directly to an article's page in GitHub for editing.
Utilizing the key findings from the survey, the competitive analysis, and the SEO audit on Bundler's site, I made the following changes to the docs page:
Rewrote the documentation landing page in order to make content more discoverable. I crafted short descriptions underneath each header to better define the different types of documentation for end users.
I focused ongoing doc maintenance efforts to v1.16 and v1.15, which were the most current releases at the time the audit was conducted.
I made the guide titles more descriptive, employing a titling scheme that would mirror search queries (i.e. “How to [do something] with[technology]”).
I worked with one of our developers to add calls to action on the footer of all of our "Guides" pages to encourage users to edit docs or open up issues in our issue tracker.
In the months following the audit (Mid-Jan 2018-May 2018), we observed the following:
An 8.38% increase in total pageviews over the previous period (Sept 2017-Early-Jan 2018).
A 7.17% increase in total users finding the pages via organic search over the previous period.
A 7.78% increase in average time on page over the previous period.