FarmStack Retro 2021-22

 Overview

 

Date

Apr 26, 2022

Team

FarmStack , Strategy

Participants

@Razak K M @Gautam Mandewalker @Waseem @Mahesh Kumar Pradhan (Unlicensed) @Rubesh @chandrashekar M S @rikin (Unlicensed) @Andrew Hicks (Unlicensed) @Shreya Agarwal (Unlicensed) @alesha (Unlicensed)

 

Year Gone By - Review

What we did well

What could have been better

Realisation

What we did well

What could have been better

Realisation

  1. No. of customer interactions

  2. New opportunities - deeper interactions in India/ Global

  3. Kenya (KALRO) onboarding

  4. Website/ presentations

  5. Data discovery tool

  6. Steward interim release

  7. FS participant mock release

  8. FAO/ UC Davis use case (outside regular projects)

  9. Indian Govt interactions

  10. TARA use case

  11. Consent collection UI: TARA and E-mircha use case

  12. Legal documents first set

  13. We did some pivots and managed well some not well-defined scenarios

  14. Documenting and re-using materials (slides, videos, writeups) - we created quite a bit of collateral this year

  1. Speed of some features in FS participant

  2. Product design of FS Participant V1

  3. Speed of implementing use cases for FAO/ UC Davis

  4. Interactions on confidential compute

  5. Managed connector with openTEAM

  6. Outcome of Agrimetrics engagement

  7. Working partnerships with HPE/ Microsoft

  8. Publicity: coherent marketing effort

  9. Fertiliser recommendation use case

  10. Clear communication of FS steward as an independent product

  11. Ontology exercise - outcome

  12. ON pilot use case - clear requirement

  13. More clearly defined goals / desired outcomes for customer interactions

  14. Communication of new features or tools with the wider team

  1. Clients have vague problem definition and it takes time to define the same and that’s the most time taking step

  2. Self-hosted connector is good but there have at least been three instances of demand for managed connector

  3. Creating even simple connector with all the containerised application still looks tedious - people want templates that can be configured

  4. Usage policy is eye-catching but nobody is able to define very well how they want to use that - that’s not the main pull for people

  5. Confidential compute is still some time away and right now govt is fine with VPC

  6. Steward has product lifecycle independent of participant (connectors) which is independent of consent/ data wallet

  7. We need to up effort for adoption and the different initiatives are:

    1. Setting up stewards and bootstrapping network

    2. Building features that bring convenience over security like templatized/ managed connectors

    3. Creating use cases with more partners - but with as little effort on engineering side as possible

    4. Aggressively enter partnerships and create example Steward to show that product is live

    5. Doing more marketing to create SEO enabled dataset explorer, newsletters, social media posts etc

  8. Need the ability to demo the software in partner calls, meetings, etc (difficult to explain via powerpoint alone)

Retrospective - Detailed

Streams

Start doing

Stop doing

Keep doing

Streams

Start doing

Stop doing

Keep doing

Engineering

  • Develop templatized connectors like GSheet/ MySQL as prioritised/ designed

  • Plan 2 week dev sprints aligned with release plan (backlogs, bugs, feature requests - designs)

  • Bring DevOps practice of fortnightly integration to the main branch

  • Build use case using participant UI with visibility in DG steward instance wherever required

  • Plan core + deployment with one squad and development with another squad

  • Monthly retros

  • Commit to implementation without complete understanding of user stories and design

  • Start implementation without detailing out the set of steps

  • Improper task description/ allocation resulting in lack of visibility of who’s working on what

  • Commit to open ended or vaguely defined problem statements

  • Sit on a bottleneck without specifying the root cause of it

  • Improve single player mode features like dataset explorer/ test bed

  • Improve designs of FS steward/ FS Participant baseline release

  • Support internal use cases and associated custom connectors

  • Continue building consent management backend

Product

  • Processes for user stories → design sprint → design review → features → dev sprint

  • Template docs for release plans, sprints, user stories, PRDs/ epics, architecture

  • Communications/ marketing, business development strategy plans

  • Onboarding/ setup guides

  • Requirement driven features

  • Lookout for business model testing

  • Deployment plan for use cases

  • Quarterly loom video updates

  • Customer funnel and interaction process

  • Monthly Retros

  • Bad Jira tickets and random confluence usage

  • Push for features without clear design review for each user story

  • Weekly sync ups

  • Customer interactions

  • Internal coordination

Product support

  • Revamp presentations

  • Supporting resources (flyers/ use case studies)

  • Process for assessing opportunities

  • Prioritized communications plan with measures of success and tools

    • Newsletter templates

    • Blog post templates

    • Social media templates

    • Plan out marketing spends as per release

    • Revisit listserv and email blasts

  • Clearer delineation of R&R in this area, who are the doers, contributors, decision makers

  • Force fit FS - if it is not required, it is not required

  • Commitment to work stream which are not clearly understood by the team ( eg, Agrimetrics)

Business Development/ Partnerships

  • Process to reach and onboard new clients

  • Onboard partners as quickly as possible (fill the wall)

  • Map out people and organisations and meeting agenda items for next quarter

  • Settle on a more clear way to communicate the FS value proposition (together with the product team)

  • Lengthy MoUs or too long time to reach a state of action

  • Using new partnership opps (like USAID bids) to share FS with a wider audience

 Action items

Action items for respective team:

@Razak K M to steer individual retros
@Razak K M to get into dev sprint planning exercise
@Gautam Mandewalker to create docs for products