Bootstrapping Test Pilot Community through Public Decision Making
This proposal is pulled verbatim from a message I sent to the Test Pilot mailing list a few minutes ago.
The question is: how do we best begin to build a community around Test Pilot?
My answer: start by making decisions in public.
If this seems interesting to you, read on below.
Proposed Participation Timeline
Q4 2016: Start by making decisions in public, in the Discourse user forums.
Q4 2016 - Q1 2017: Once we’ve made our process accessible to contributors, ask active Mozillians to get involved. Build an awesome core community. Advertise idea submissions to active Mozillians & iterate on the submission system before a huge public influx.
Q1-Q2 2017: Once the core community is in place, and idea submission has been tweaked, get Marketing involved with a public call for ideas. Our awesome core community will help manage the influx: greet newcomers, bash trolls, de-dupe suggestions.
To frame the discussion, I wrote up some thoughts on what a more open, community-centered product development cycle might look like. TL;DR: give community a seat at the table by offering equal visibility into the process, and opportunities to provide input at decision points.
See also Producing OSS, which makes interesting points on the importance of public decision-making.
Suggested Q4 plan
In Q4, we can set the stage for community by making our work public:
- ask product and UX to move decision making discussions to Discourse
- ask experiment teams to post in Discourse as they launch, measure, iterate, and wind down
- make Discourse a secondary call to action on the Test Pilot experiment pages
- deprecate mailing list in favor of Discourse
- (if there’s time) ask vetted, active Mozillians to join the conversation
- (if there’s time) ask Mozillians to share ideas
If we make these changes in Q4, then by Q1, we’ll have plenty of content in Discourse. New community members will tend to model their input on the tone and content of the existing discussion; for this reason, I think we should hold the open call until a bit later. I realize I’m contradicting my own recent suggestion, but I do think this approach will yield better results.
Key Result: 100 monthly active Discourse users
One natural number to measure would be the number of monthly active Discourse users (MADU), meaning: users who post at least once a month in the Test Pilot category. Right now, this number is probably below 10. If all the dev teams and product/UX get involved, that’ll jump to, maybe, 30 MADUs. If we get Mozillians engaged, we could see this number jump into the hundreds. 100 MADUs seems ambitious, but possible.
Q4 plan in detail
It’s a bit buried in the list above, but I think we should deprecate the mailing list, and move discussion to Discourse instead. Our current list has little traffic, and a primitive, plain text archive that doesn’t allow search. It would be easy to move the current traffic to Discourse. Finally, moving to Discourse would encourage the Test Pilot team to use it.
Going into a bit more detail on the types of content that should move to Discourse to help seed our community:
Design phase (product / UX):
- Idea proposals & discussion
- Decision making discussions: which ideas to invest in, and why those ideas.
Development / iteration phase (product / UX / dev):
- Summaries of user interaction data (how are people using the product, how many keep using it)
- High level discussion of which features to add, change, or remove, grounded in the public interaction data
- I’m not yet sure how much development discussion should happen on Discourse vs. a specific mailing list; we can ask around to see what approaches different teams would like to try.
Graduation / end of cycle phase (product):
- Discussions about whether to keep an experiment running, move into Firefox, or retire it.
That does it for the mailing list post.
What do you think about this proposal? Is this a good way to build the foundations of a participatory Test Pilot community?
Let me know! Twitter and email are in the footer below.