A developer platform lives and dies by it's developer community. When huge problems need to be solved, it's easy to make valuable improvements, but what do you do when those are solved and you still see high bounce rates on your site, low developer application completion, and generally poor adoption of your product? This is where your data can save you.
In this talk we'll run through:
- How to track valuable developer path insights, from moments of anxiety to time to first valuable call.
- Overlaying support and ticketing information on top of developer path data to decrease developer friction.
- How to create automated analytics systems to measure success.
- When these systems should be built, before it's too late.
2. Why is the Data Important?
When you have big problems, solutions
are easy. When those are solved how do
you determine what will make any
measurable impact?
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
3. Agenda for Today
Identifying Hurdles: Using pull based data analytics
to identify developer hurdles.
Improving Transparency: Using pushed based data
systems to decrease churn and reduce ticket volume.
Improving Onboarding: Using this data in practice.
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
5. 1 Identifying key moments of anxiety
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
6. 1 How it translates to developer onboarding
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
7. 1 What metrics are we looking for?
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Time to first call
How long did it take the developer to make
their first valuable API call?
Developer journey vs support
Which pages / paths did the user follow,
which pages did they leave on, and what
support requests did they file?
8. 1 How do we obtain these metrics?
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Determine onboarding channels: What systems will a developer go through to
make their first API call? This includes support channels.
Unify the tracking identifiers: You need a way of tracking the non logged in user
uniformly through these channels, such as a tracking ID.
Remove edge / non-valuable data: Single path occurrences are outliers and
should be ignored. Valuable time to first call metrics should remove auth
requests and any test samples that skew the data set.
Don’t lead your data: If you have to makes leaps based on your anecdotal
information to come to a conclusion, you’re missing a channel. Anecdotal
information should only help bolster a data proven conclusion.
9. 1 Enhance with Anecdotal and Support Information
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Time to first call
Support tickets, feedback / community systems, and
anecdotal customer support requests are typically
major areas to determine methods for decreasing
time to first call (e.g. improving auth docs)
Developer journey vs support
Customer feedback, major areas of support
requests, and direct ticket request information
are all prime supporting information for
Improving developer onboarding and products.
11. 2 Roadmap Transparency
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
12. 2 What should the roadmap focus on
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Avoid exact timelines: Using terminology to denote near, medium,
and long term targets avoids being seen as unreliable if dates slip.
Avoid roadmap changes: Only display what you can commit to, such
as a 3-6 month timeframe.
Use it as an engagement mechanism: Collect feedback on what
people want / don’t want on the roadmap, then use that to improve
it.
13. 2 Changelog / Release Notes
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
14. 2 What should the changelog include?
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Granular scoping: Ability to filter changes based on relevant topic,
such as API changes, breaking changes, particular product lines, etc.
Subscription: Treat it as an independent service, allow users to
subscribe to updates via RSS feed, email, text, etc.
Overshare: Breaking changes, enhancements, upcoming major
changes, maintenance, and the like should be shared. This isn’t a
newsletter where you’re worried to communicate too often.
15. 2 API Uptime Status Indicator
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
16. 2 What should the status indicator include?
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Granular statuses: Ability to filter changes based on relevant topic,
such as API changes, breaking changes, particular product lines, etc.
Subscription: Treat it as an independent service, allow users to
subscribe to status changes for any granular area of the APIs /
services via RSS feed, email, text, etc.
Immediate and continuous posting: When service changes occur,
they need to be reflected immediately. Status updates should be
posted regularly until problems are rectified.
18. 3 Using the Data: Documentation Rebuild Example
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
19. 3 Using the Data: Documentation Rebuild Example
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Source: Box developer site at https://developer.box.com
Core data sources: User site path, time on page, exit rates, and bounce rates.
Supporting sources: Support ticket data, forum feedback, direct customer
feedback, and independent audit.
Resulting improvement:
Page view increase: 2x
Bounce rate decrease: 68.68% -> 24.64%
Exit rate decrease: 60.12% -> 24.19%
20. 3 Improving Transparency Example
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
21. 3
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com
Source: Box Pulse at https://pulse.box.com
Core driving factor: Direct customer feedback, transparency improvements
Requirements: All feedback is under strict SLAs, all teams responsible for
responding and assigning a status to feedback within a given period of time.
Resulting improvement: 2834 independent pieces of feedback, 225 delivered,
332 being researched, 247 under consideration, 7 in beta, and 74 on the
roadmap. All within 3-4 months of being launched.
Improving Transparency Example
22. Wrapping up our topics
Identifying Hurdles: Using pull based data analytics
to identify developer hurdles.
Improving Transparency: Using pushed based data
systems to decrease churn and reduce ticket volume.
Improving Onboarding: Using this data in practice.
Jonathan LeBlanc • Director of Developer Advocacy @ Box • Twitter: @jcleblanc • Email: jleblanc@box.com