How I Think About Sprint Planning as an Engineering Manager
Five questions I answer before every sprint, and why planning for real capacity, including training and production support, builds trust.
Before a sprint starts, I like answering five questions. They’re simple, but they catch a surprising number of problems before they turn into missed commitments.
- What business outcome are we trying to achieve?
- What work must be completed during this sprint?
- Who owns each deliverable?
- What dependencies could block us?
- What capacity does the team realistically have?
Let me walk through why each one matters.
1. What business outcome are we trying to achieve?
A sprint shouldn’t just be a container of tickets. It should move something forward. When the team knows the outcome, like “establish the foundation for the new platform” or “fix the data quality issues affecting lead reporting,” it can make smarter trade-offs when something unexpected happens mid-sprint.
2. What work must be completed?
There’s a difference between work we’d like to finish and work that must finish. I want the team to know which items are essential to the outcome, so that if capacity shrinks, we protect the right things.
3. Who owns each deliverable?
Unowned work drifts. “The team” owning something often means nobody is watching it. Every meaningful deliverable should have a named owner, even if several people contribute.
4. What dependencies could block us?
Dependencies are the most common reason good plans fail. Waiting on an approval, an API from another team, a vendor, an environment, or content from stakeholders can stall work that’s otherwise ready. Identifying these at planning time lets us start the conversations early, or sequence work so we’re not stuck.
5. What capacity does the team realistically have?
That last question is the one I spend the most time on, because it’s the one most often answered optimistically.
Engineers are not working in a vacuum:
- Production issues happen.
- Meetings happen.
- People take PTO.
- Training is necessary.
- Dependencies slip.
- Code review and QA take real time.
A plan that ignores these things isn’t a plan. It’s a hope. And when it fails, it’s usually the team that gets blamed for “missing the sprint,” even though the plan was never realistic.
Planning for learning, not around it
One example from my own planning makes this concrete.
During a major platform migration, I intentionally account for training alongside migration work, rather than pretending engineers can learn an entirely new technology stack outside the sprint.
In our current planning, JavaScript, TypeScript, React and Next.js learning are incorporated alongside the migration phases rather than treated as unrelated activities. Learning is on the plan. It has time allocated. It’s visible to stakeholders.
Why does this matter? Because the alternative is a hidden tax. If training isn’t planned, engineers either learn on nights and weekends, which isn’t sustainable, or they learn during the sprint without it being accounted for, which quietly reduces capacity and makes the team look slower than planned. Either way, the plan is wrong.
When learning is part of the plan, the numbers are honest, and the team gets the time it actually needs.
Realistic planning builds trust
Stakeholders don’t need a team that promises everything. They need a team whose promises they can rely on.
When sprint planning reflects real capacity, a few good things happen:
- Commitments are met more consistently.
- Surprises are discovered at planning time instead of at the sprint review.
- Stakeholders get honest timelines, so they can plan their own work.
- The team stops carrying the stress of a plan they knew was unrealistic.
Realistic sprint planning builds trust, sprint after sprint. And trust is what gives an engineering team the credibility to push back, negotiate scope and ask for the time to do things properly.