How does a feature or product come to be? Using the new feature, Special:Insights, as an example, Stanley and Trevor take us through the process.

How do ideas become features?

  • Scientific Method used to determine if a feature is viable
    • Gathering a hypothesis, gather data, analyze data
  • Product development
    • Ideas become prioritized and developed

Product Development Cycle

  • Ideas
  • Prioritization
  • Research, Design & Definition
  • Development
  • Feature Release
  • Follow-up
  • Repeat!

Where do we get our Ideas?

  • Problems in the current Wikia product
  • Our users
  • New technology
  • Wikia employees
  • Inspiring Products
  • Market Trends

What inspired the Idea for Special:Insights?

  • Users wanted a better platform to view various stats/data
  • Special:WikiStats was limited and considered “technical debt”
  • Other major websites offer their users comprehensive data


Allocating time on projects can difficult to balance with employees times.

  • prioritization in a nutshell; “dedicating employee time on _____ will best help us achieve our goals”
    • Special:Insights Example - “Dedicating employee time on surfacing site data to contributors will best help us achieve more complete article pages.”
      • Why? The most requested feature from our active users is data.


  • Sources for research
  • Surveys
  • A/B tests
  • Discussions with users
  • User testing
  • Industry analysis
    • Special:Insights Example(s) -
      • What are the trends of our users?
      • What do other sites offer their users?
      • What kind of data do our users want?


  • Designs have many drafts and iterations. They evolve and adapt according to the research stage.
  • What to consider when designing:
    • Audience
    • Feasibility
    • Internationalization
    • Multi-device support
    • Consistency with the rest of our products


  • Minimum Viable Product aka MVP
    • Delivering fully functional feature or product
    • Instead of taking 1-2 years to make something that might fail, making a product with many versions for quick feedback and learning from each iteration
      • Special:Insights Example
        • Focused on making the feature encourage edits as opposed to just showing data
        • MVP was 4 lists that we could quickly get out of the door

Development & Release

  • “Fail fast” - build intelligently yet quickly so we can validate hypothesis against real users. Then iterate.
  • Multiple release vehicles available
    • One single alpha community
    • Labs
    • By languages
    • Opt-ins
      • Special:Insights Example -
        • Recycled existing special pages for quick go to market
        • Built and tested in 3 weeks
        • Shipped to a small handful of test communities


  • Collect & respond to user feedback
  • Bug fixes
  • Measure and reflection. Did we achieve our goals? Where can we improve? What are the next steps?
    • Special:Insights Example -
      • User feedback
        • Feedback has been mostly positive
        • The data is ok, but not really what I asked for. I want more!
        • This isn’t new, I already have these special pages
      • Usage data shows
        • Most users fix one or two issues back to back
        • Some users/communities have never used the feature
        • Popular issue: The lists are finite
      • Potential next steps
        • Expand on lists - Give users more relevant data
        • Explore and experiment with participation drivers