Stop Reassigning People Mid-Stream

A fast product development firm that had just joined a larger company was pulling engineers across client work, internal initiatives, and urgent requests mid-sprint. Smaller dedicated teams, planned batches of work, and a buffer for interruptions reduced the churn without giving up the firm's speed.

05 FLOW
The condition

Engineers were being moved between projects in the middle of a sprint, and every move looked like responsiveness.

Speed That Now Had to Answer Upward

A product development firm had built its reputation on rapid iteration, taking medical device, health technology, and sports technology products from idea to working prototype quickly. Then it joined a larger corporate organization through acquisition. Its speed now had to live alongside longer planning horizons, broader priorities, and resource commitments the parent company could see.

The firm's leaders were clear about what they wanted from the change: stronger strategic alignment, more consistent delivery, long-term visibility into resources and progress, and less context switching across projects. They also did not want to lose the speed and adaptability that made them valuable. The last two goals pulled against each other, and the way the firm allocated its engineers sat right between them.

Focus Was Treated as Spare Capacity

Delivery ran through a single team of about twenty people supporting a wide range of product initiatives. Every planning conversation ran long, and every decision took longer than it should have. The same scarce specialists were pulled into request after request.

Engineers were reassigned across projects in the middle of a sprint: client work, internal initiatives, urgent requests. Each move was a reasonable answer to the request in front of someone. Together, the moves created churn and made the team's forecasts a little harder to believe each time.

That is the trap in mid-stream reassignment. Every individual move looks like good resource management, and none of them carries its full cost on the day it is made. The context an engineer leaves behind, the work that slows when they go, and the time the receiving work spends absorbing them never show up as a line item. Nobody had decided that focus was optional. It had simply become the first thing spent whenever something urgent arrived, because nothing in how the work was planned protected it.

Figure 1
How engineers were allocated
Reassign when asked
  • One team of about twenty
  • Engineers moved mid-sprint
  • Specialists on every request
  • Urgent work takes anyone
Plan for focus
  • Smaller domain teams
  • Dedicated, little rotation
  • Scarce skills mapped ahead
  • A buffer for urgent work
The old model let each urgent request pick its engineer. The new one planned where people and scarce skills would be before the requests arrived.

Smaller Teams Built Around Domains

The first change was to the team itself. The single large team was replaced by smaller teams aligned to specific domains and components of the product, such as hardware, firmware, and research and development. Each had its own Product Owner and Project Engineer. Instead of the whole group meeting to stay in sync, team leads coordinated lightly with one another.

Leaders from the parent company joined the teams as Product Owners, so priorities reached the builders directly instead of through layers of people. Above the existing two-week sprints, the firm added quarterly planning. Teams could now see longer-term initiatives, spot dependencies across domains earlier, and make capacity trade-offs with the bigger picture in view instead of one sprint at a time.

Figure 2
One large team, regrouped by domain
About 20 people on one team
  1. Hardware
  2. Firmware
  3. Research and development
The single team of about twenty was split into smaller teams aligned to parts of the product. The three domains shown are the examples the firm gave, not a complete list.
Two scientists collaborating over samples and evidence in a modern laboratory.

A Buffer Instead of a Fire Drill

The second change went directly at reassignment. A new resource planning model emphasized dedicated teams, work planned in batches ahead of time, and minimal rotation. Teams planned further ahead and mapped where scarce skills would be needed before the crunch, rather than borrowing them when it arrived.

Planning became capacity-informed: teams committed against the capacity they actually had. And a buffer was set aside for urgent requests, so new demand did not have to pull an engineer off their sprint every time it appeared. The buffer did not pretend interruptions would stop. It gave them a place to land that had been planned for.

Steadier Sprints, With Honest Caveats

Teams reported less mid-sprint churn and more confidence in what they would deliver. Engineers stayed more focused, and delivery velocity improved through better prioritization and more consistent task closure. Some teams approached 85 percent completion early in the quarterly planning cycle. Results varied, because vacations and outside dependencies did not pause for the new model.

The smaller teams changed the meeting load as well. Sprint planning and refinement meetings got 25 to 50 percent shorter for the teams involved, and cross-team dependencies were flagged and resolved earlier. Not everything improved evenly. Quarterly planning raised the quality of planning, but its time savings were not consistently realized, and work that crossed teams or newly identified priorities still occasionally disrupted plans.

The firm's next steps were to extend quarterly planning to more teams, make more use of performance metrics, share practices across domains, and standardize meeting cadences. None of it required giving up speed. It required treating an engineer's focus as something to plan for, rather than the reserve every new request could draw on.

Figure 3
What changed, with its limits
About 20people on one team beforeReplaced by domain teams
25 to 50%shorter planning meetingsTeams involved
Near 85%done early in the quarterSome teams; results varied
Early results from the teams involved, not a controlled study. Completion near 85 percent applied to some teams, and vacations and outside dependencies kept results uneven.

Interruptions arrive either way. The only choice is whether they are planned for or paid for by whoever gets pulled off their work.

Share With Your Team

Choose who it is for. You send it yourself from your own email; Agileon never writes to your colleagues.

Not sure where to start?

Answer five questions about where work stalls. We send back a starting plan: where to begin and what to look at first.