The Expedite Lane That Ate the Platform Team
Helios Cloud's platform engineering group has 14 engineers supporting 30 product teams. Its Kanban board has five columns and an Expedite lane capped at three items - but on most days eight to ten items sit there. Two stakeholders, the head of security and the VP of sales, routinely flag items as expedite; together they account for 70% of expedite entries. Incident response consumes 55–65% of capacity in any given week, varying with releases. The product organization has committed a six-month roadmap that assumes 80% planned capacity.
Standard-class items average 27 days from start to finish; expedited items average 3. The team is exhausted; two engineers have resigned. The director of platform has hired you to make the board work and has asked for a new expedite policy by the end of the month.
1. Is this primarily a board design problem? What would you investigate before proposing policy?
Read the worked answer
This is not primarily a board design problem. The expedite lane is doing what it was built to do: absorbing unmanaged demand. Eight to ten items in a three-item lane, two stakeholders generating 70% of entries, and 55–65% of capacity consumed by incidents together point to a demand and decision-rights problem.
Before proposing policy I would investigate four things. First, demand: an eight-to-twelve-week Pareto of all incoming work by source, type, and whether it was genuinely time-critical. Second, incidents: how many are repeat failures, and what upstream defect or design issue drives them; if incidents recur, expediting them faster is the wrong fix. Third, the capacity assumption: the roadmap assumes 80% planned capacity while the data shows 55–65% unplanned, so the plan is already wrong and will keep generating pressure. Fourth, decision rights: who can declare expedite, on what criteria, and with what consequence.
I would also check whether the two stakeholders' requests are truly expedite or simply important - security and sales often have real deadline pressure that a fixed-date or fast-track class can serve better. The output of the investigation is a shared picture of trade-offs, not a policy document written by the coach.
2. Design a coaching process to produce an expedite policy that the security lead, sales VP, director, and team will actually follow.
Read the worked answer
Policy written by the coach will be ignored; policy written by the people who hold the pressure has a chance. I would run a structured workshop with the director, head of security, VP of sales, two engineers, and the incident lead - not delegates. Pre-work: the demand data, the 27-day versus 3-day comparison, and a one-page description of four classes of service with proposed entry criteria.
The workshop sequence: agree what expedite means (time-critical, cannot wait for the normal sequence); quantify the cost of each expedite in displaced standard work using the data; agree capacity allocation - for example, a reserved band for unplanned work, a cap on concurrent expedites, and a named decision owner with a rotation; define a fixed-date class for sales commitments and an SLE-based standard class for recurring security requests. Then agree the review cadence and what triggers a policy change.
I would coach the stakeholders through the trade-off rather than arbitrate it: the sales VP needs to see that every expedite moves a standard item 27 days, and the security lead needs a predictable route that does not require flagging everything. The director must own the policy publicly. A policy that the director can override informally is not a policy, and the first test will come within a week.
3. Two engineers resigned and the team is exhausted. What is your duty as coach, and how does it change your sequencing?
Read the worked answer
Resignations and exhaustion are system outcomes, and my duty is to name them as a business risk to the sponsor rather than treat them as engagement friction. Fourteen engineers absorbing 55–65% unplanned work with an eight-to-ten item expedite queue is not sustainable, and a new policy that simply routes work faster will accelerate the loss.
Concretely, I would raise with the director: the two resignations, overtime and after-hours incident patterns, and the mismatch between the 80% planned capacity assumption and the actual demand mix. I would recommend a short moratorium on new roadmap commitments until capacity allocation is agreed, or at minimum an explicit statement to product that committed dates carry a stated confidence level.
Sequencing changes: capacity allocation and WIP reduction come before board optimization or expedite triage. I would also protect the team's visibility - make workload visible to stakeholders without exposing individuals. If the sponsor rejects any relief and wants only faster expediting, I would say plainly that I cannot coach toward that outcome, document the advice, and let the sponsor decide. Sustainable pace is a fitness criterion, not a wellness extra; higher-level practices do not survive on top of it.