Predictable delivery against a fixed government deadline comes from committing honestly rather than working faster: plan each sprint from days-off-adjusted capacity, hold back a measured buffer for unplanned work, and track the say-do ratio — the share of committed work actually delivered — so the must-have scope burns down on schedule instead of on hope.
A state benefits portal has to be live before open enrollment on November 1. The date sits in statute and in the governor’s press release, and it is not moving. A hiring freeze means the team’s two vacancies stay vacant through the fiscal year, and part of the group still carries production support for the legacy system. Fewer people, same mission, same date. Programs in this position rarely fail for lack of effort — they fail because every sprint was planned as if none of this were true.
Why do fixed-deadline programs miss dates even with capable teams?
The failure is rarely one catastrophic slip. It is a slow leak: the team commits to what the roadmap needs instead of what the roster supports, misses by fifteen or twenty percent, rolls the remainder into the next sprint, and commits high again. Each individual miss looks recoverable. Compounded over a dozen sprints, the program arrives at its final quarter needing a velocity it has never once demonstrated.
Three quiet forces drive the leak:
- Unplanned work treated as an exception. Production incidents, security patches, and data calls from oversight bodies arrive every sprint, yet most sprint plans assume zero of them.
- Days off discovered mid-sprint. Holidays, training, and leave are knowable at planning time but are usually averaged away — until the sprint where they cluster.
- Commitments anchored to hope. When the deadline is fixed, there is pressure to commit to the schedule’s number rather than the team’s number.
In a commercial team, a missed sprint is a private disappointment. In an agency, the commitment travels up a chain — scrum master to program manager to program office to oversight — and every miss spends credibility the program will need later, when it has to ask for a scope decision or more people.
How do you make sprint commitments predictable?
Five practices, applied every sprint rather than once at kickoff:
- Separate must-have from should-have scope. Statutory and mandated items go on one line, enhancements on another. The burn-up that decides the program’s fate is the must-have line against the fixed date; track it separately.
- Start from demonstrated velocity, then adjust for the calendar. Take the trailing three-to-five-sprint average of completed points. For each upcoming sprint, subtract holidays, approved leave, training days, and detail assignments person by person, and scale the velocity by the share of person-days actually available.
- Measure unplanned work instead of guessing it. Look at the last three or four sprints and compute the share of capacity that went to work nobody planned. Reserve that share explicitly in every future sprint.
- Commit below adjusted capacity and track say-do. The commitment is adjusted velocity minus the unplanned-work reserve. The say-do ratio — points delivered divided by points committed — gets reviewed at every retrospective, not just at milestones.
- Escalate on trend, not on crisis. If the must-have burn-up runs below the required line for two consecutive sprints, raise it with the numbers while descoping, borrowing capacity, or renegotiating scope are still real options.
A worked example: six sprints to open enrollment
An eight-person state team runs two-week sprints, with six sprints left before the November 1 cutover. Trailing four-sprint velocity averages 32 points. Over the last three sprints the team committed 36, 38, and 35 points of planned work and completed 30, 33, and 31 — a say-do ratio around 86%, with the shortfall rolling forward every time.
The look-back explains the gap: roughly 15% of each sprint went to unplanned work — legacy production support and monthly security patching — that never appeared in the plan. Plannable throughput is therefore about 32 × 0.85 ≈ 27 points. Sprint 3 additionally contains a federal holiday and mandatory training for two staff, removing 9 of 80 person-days; that sprint scales to roughly 28 points of adjusted velocity and 24 committable.
Predictable capacity to the deadline: five sprints × 27 points plus one reduced sprint at 24 points = 159 points. The must-have backlog stands at 170. The 11-point gap is visible in August, while it is still a planning decision — defer two enhancement stories out of the must-have window, or borrow a developer from a sister team for two sprints — rather than an October crisis. Just as important, a team that commits 27 and delivers 27 produces something the old commit-36-deliver-31 pattern never could: a forecast leadership can take upstairs without hedging.
Making the discipline repeatable in Jira
None of the arithmetic is hard. What is hard is doing it every two weeks when velocity lives in one report, leave in a shared calendar, allocation in a staffing spreadsheet, and the commitment in someone’s meeting notes. Sprint Planning with Capacity Planning for Jira puts past velocity, each person’s days off, and allocation beside the backlog on one screen, so the commit-below-capacity check happens during planning rather than after it. Each sprint’s plan then doubles as a record — who was allocated, which days off were subtracted, why the team committed to the number it did — the kind of decision trail program offices and oversight reviews increasingly expect to see. For agencies consolidating tools as they move to Atlassian Cloud or Atlassian Government Cloud (AGC), that discipline arrives inside the Jira backlog instead of as another system to maintain. The app is Cloud Fortified and runs on Atlassian’s SOC 2 Type II / ISO 27001 platform.
FAQ
What is a good say-do ratio for a government agile team?
Sustained 90% or better. Below about 85%, forecasts built on the team’s commitments stop being credible, and the commitment size — not the team — is usually the problem. A permanent 100%, though, can signal under-commitment; occasional small misses are normal.
Is committing below capacity just sandbagging?
No. The reserve is not idle time — it absorbs the unplanned work that arrives every sprint whether planned for or not. Making it explicit is more honest than absorbing it silently, and if a quiet sprint leaves the reserve unused, the team simply pulls the next must-have item forward. Nothing is lost.
How big should the unplanned-work buffer be?
Measure it — three or four sprints of history is enough. Government teams commonly land between 10% and 25%. If the measured share exceeds roughly 30%, the unplanned demand itself is the problem to manage down, not something to buffer around.
A practical first step before your next planning session: pull the last four sprints and compute the say-do ratio. If it is under 90%, set the next commitment to your delivered average adjusted for that sprint’s days off — and write down why. Two sprints of hitting exactly what you commit will do more for your November credibility than any status-report narrative.
Try it: Install free on the Atlassian Marketplace · Help docs
Leave a Reply
Your email is safe with us.