2018-12-13 Retrospective - Community Portals

2018-12-13 Retrospective - Community Portals

Date

Dec 13, 2018

Participants

@Meredith Slota (Unlicensed) @Michael Lee (Unlicensed) @Michael Barakat (Unlicensed) @Anna Greenwood (Unlicensed) @Jay Hodgson @Sara Gosline (Unlicensed) @Robert Allaway @Ljubomir Bradic (Unlicensed) 

Background

The "community portal" product team is a cross-functional group with members from the platform and scientific teams. After working on two portal prototypes for two months, and shipping versions of both live, it seems useful to check in and examine our working practices and see where we need to make tweaks moving forward. 

Retrospective

Whiteboard photos at the bottom; below are transcribed notes, followed by a grouping into start / stop / continue + relevant Jira tickets as action items. 

What Worked

Launched! On time! Deadlines met!

Issues were filed (Jira)

Issue clarity was ok

Flexibility - chance coure, iterative

Cautious with feature creep despite speed

Realistic deadlines / deliverables / scope

 Kept the customer in mind

group_concat shipped

 ANONYMOUS ACCESS TO DATA 

Priorities:

  • Everyone had a voice, was a leader

  • Good listening

  • Good communication, triage

  • Good discussions

No overreach in scope

"Not under the gun"

Clear ownership / shared responsibility

Clarified was "next level" CMS might be: wikis > structured content > schemas

Dogfood worked! Using "explore" to find errors in data

What Could Be Better

Jira workflows - validation / closed

  • Comments used for status

  • Confusion over "shipping"

  • What is "done"?

What are "versions"? - dates for v1, v2 NF are unclear

 NO ROADMAP BY SPRINT

  • design vs. dev happening / overlapping

  • Gantt chart for project? 

  • difficult to estimate LOE + dev time

NO TIME FOR FOUNDATIONAL RESEARCH

  • what even is a data portal for? 

  • tools / products not aligned to key scenarios

Difficult to cut "explore" from NF

Not owning the story

  • Product messaging, no communication channels externally

  • Partial project for AMP-AD caused confusion

Ownership between Anna / Meredith - some gaps, some overlap

Didn't walk through whole user process / main scenarios

  • Uncovered technical risk re: tables / download LATE

  • Who owns QA? unclear and confusing

BIG PICTURE IS UNCLEAR - how do portals + Synapse work? 

  • Messaging to internal folks vs. partners

Too much power given to partners (e.g. URLs, naming)

No documentation of discussions causing churn

Design changed over time but files disorganized, out-of-date

  • No single source of truth about what we're building

SPF for components - risk and stressor

  • Want more time for testing

  • Artificial divide between MB + Synapse engineers

Lack of focus on stabilizing components

Not enough wiggle room in ship dates

Start doing

Stop doing

Keep doing

Start doing

Stop doing

Keep doing

  • Coordinated shipping process

  • Consistent QA testing

  • Clearing technical debt

  • Sprint planning?! 

  • Document relevant decisions

  • Clarify versions + dates

  • Long-term roadmap for portals AND Synapse

  • Inflexible ship dates

  • Making decisions without documenting them

  • Treating portals as totally separate from Synapse

  • Assuming we know what users want  

  • Jira works if you work it!

  • Stand-ups saved tons of time

  • Cross-functional engagement

  • Design + dev in the same room

  • Clear product ownership