Skip to content

Latest commit

 

History

History
183 lines (117 loc) · 8.08 KB

File metadata and controls

183 lines (117 loc) · 8.08 KB

The bimstack sprint

Listen → Map → Make → Test → Ship → Show → Iterate

Seven verbs. Every sprint, every phase, every service. Not a process to follow in order – a set of activities a team is always rotating through.

This file explains what each verb means in practice, who leads it, and what good looks like.


Listen

What it is: every activity that brings the citizen's voice into the team. Not just user research in a discovery; continuous listening throughout the life of the service.

Who leads: service designer, supported by content & interaction designer.

What good looks like:

  • At least one round of user research every sprint
  • Front-line staff present in show-and-tells
  • Citizen quotes in weeknotes, by situation (anonymised if needed)
  • Existing data (analytics, support tickets) reviewed every sprint
  • Under-represented citizens specifically sought out, not just whoever's easiest to recruit

Skills: bimstack:research-coach (the router across the whole cycle), bimstack:research-planning, bimstack:transcript-analysis, bimstack:research-presenting, supported by design:user-research and design:research-synthesis.

Cost of skipping: you build for an imagined citizen. Standard 1 fails.


Map

What it is: turning what you've heard into a shared view of the service – the journey, the actors, the systems, the touchpoints. Maps are how a multidisciplinary team agrees what they're talking about.

Who leads: service designer.

What good looks like:

  • Current-state journey before any future-state design
  • Ecosystem map identifying every actor and system, not just the digital ones
  • Service blueprint making the back office visible from beta onwards
  • All maps published to the team space (and outside the team where possible)

Skills: bimstack:journey-map, bimstack:service-blueprint, bimstack:ecosystem-map, bimstack:experience-map (the Anthropic service-design plugin extends these where installed).

Cost of skipping: the team builds different services in parallel and doesn't notice until beta.


Make

What it is: turning the map into something a citizen can use. In alpha, that's clickable prototypes. In beta, it's production code. In live, it's iterations on production.

Who leads: content & interaction designer for early prototypes; developer once production code lands.

What good looks like:

  • Multiple alphas in alpha (not one)
  • Prototypes use the Barbados Design System – no Tailwind, no bespoke colours
  • Production code uses lookups (Trident ID, vehicle, business), not retyped fields
  • Page weight kept low – tested on a real phone, on a slow connection
  • Source code in a public repository from day one (Standard 9)

Skills: anthropic-skills:govtech-barbados-services, anthropic-skills:govtech-barbados-forms, design:design-system.

Cost of skipping: you have ideas instead of evidence.


Test

What it is: putting the made thing in front of people and systems that will break it – users, assistive tech, slow networks, attackers.

Who leads: content & interaction designer for usability and accessibility; developer for performance; cyber engineer for security.

What good looks like:

  • Usability testing every sprint, with real users on real phones
  • Accessibility audit every sprint, including with assistive tech
  • Performance testing on throttled connections before any release
  • Security testing including a pen test before public beta, and at least annually thereafter
  • Bugs and friction are filed in the open

Skills: bimstack:synthetic-research, design:accessibility-review, design:design-critique.

Cost of skipping: the citizen tests it in production, and Standard 5 fails.


Ship

What it is: putting the thing in front of real citizens, in production, in small enough chunks to keep the risk small. Discovery → alpha → beta → live; private beta → public beta; small frequent releases.

Who leads: delivery manager, supported by developer.

What good looks like:

  • Private beta before public beta, always
  • Small, frequent releases – not big-bang launches
  • Operational readiness in place – on-call, incident runbook, support process
  • The four GDS baseline metrics wired up before public beta
  • Phase-gate standards assessment before every transition

Skills: bimstack:service-standard-assessment.

Cost of skipping: you ship a service nobody is ready to run.


Show

What it is: working in the open. Weeknotes, show-and-tells, public repos, decisions written down where the next team can find them.

Who leads: delivery manager, but everyone takes a turn at the front.

What good looks like:

  • Weeknote every week, even when the week was rough
  • Show-and-tell every sprint, with the thing shown live
  • Source in a public repository from day one
  • Decisions captured in ADRs in the repo
  • Other MDAs and civil society invited

Skills: bimstack:weeknote, bimstack:show-the-thing, anthropic-skills:govtech-barbados-presentations.

Cost of skipping: the team gets isolated, the knowledge stays in heads, and Standard 9 fails.


Iterate

What it is: keeping going. The end of one sprint is the start of the next. The end of one phase is the start of the next. Live is not maintenance – it's where most of a service's life is spent and most of the value is delivered.

Who leads: the whole team.

What good looks like:

  • Every sprint feeds back into the next sprint's plan
  • Live phase has the team and budget to make substantive changes, not just bug fixes
  • Standards reassessment at least annually in live
  • Retirement planned for when the need is gone

Skills: every skill in the bimstack, repeatedly.

Cost of skipping: the service drifts out of fit, Standard 10 fails, and citizens stop using it.


How the verbs map to the phases

Phase Listen Map Make Test Ship Show Iterate
Discovery Heavy Heavy Weeknotes, show-and-tells Through discovery rounds
Alpha Heavy Heavy Prototypes Heavy Weeknotes, show-and-tells Through alpha rounds
Beta Steady Steady (blueprint) Production Heavy Yes (private then public) Weeknotes, show-and-tells Every sprint
Live Continuous Periodic Iterative Continuous Small frequent releases Weeknotes, show-and-tells Forever, then retire

A team in alpha that's heavy on Make and light on Listen has misread the phase. A team in beta that hasn't started Test is heading for trouble.


What's deliberately not in the sprint

bimstack's sprint excludes some things that show up in lots of other agile frameworks:

  • Estimation rituals. Story points, t-shirt sizes, planning poker. They're a proxy for actual planning and rarely help.
  • Stage-gate sign-offs from outside the team. The team owns the work. The Standards self-assessment is the gate.
  • Velocity charts. Velocity is a vanity metric. Standard 13's four metrics are the real ones.
  • Big upfront design. Anything not validated with users is a guess.
  • Big upfront architecture. ADRs accumulate; they don't sit at the top of the backlog.

These aren't banned. They're just not the defaults.


What to do when a verb is being skipped

If the team is skipping a verb, name it. In the weeknote. In the show-and-tell. In the standards self-assessment.

You notice Likely cause What to do
No Listen this sprint Recruitment fell through, or research time got squeezed Block research time first in the next sprint plan
Make without Map The team jumped to a solution Pause and map. Even a 30-minute whiteboard helps.
Test only at the end Test got pushed back by a deadline Move test forward. Late testing is fake testing.
Ship without Show The team got busy Catch up. Two weeks without a weeknote is a smell.
No Iterate after Ship Treating live as maintenance Standard 10 is failing. Bring the team back in.

The sprint mantra is a check, not a recipe. If something feels off, walk the verbs and find the one that's missing.