One of the fundamental differences between traditional waterfall delivery and agile delivery is planning. For the former, planning is performed upfront with an entire view of scope and an end date, typically defined by management. For the latter, planning of performed iteratively. [1] The two common variants for iterative-based planning are capacity-based and velocity-based planning.
Capacity-based planning is based on the premise that the team selects work during Sprint Planning based on their expected capacity for that Sprint. The Product Owner brings the top-priority Product Backlog Items into the meeting and describes to the team, usually starting with an overview of the set of high-priority items. Following that, team members select work to bring into the Sprint not exceeding their capacity. [2]
The focus in determining scope using capacity-based planning relies less on Story Point estimation and more team judgement. Story Points can be used as a sanity check for both Sprint scope and estimations.
Velocity-based planning is based on the premise that the amount of work a team will do in an upcoming Sprint is roughly equal to what they have done in prior Sprints. [3] Various assumptions are made such as a constant team size, similar type of work, consistent Sprint lengths, etc.
To utilise velocity-based planning the team needs to have base line data to form assumptions around the anticipated future velocity of the team. Hence, Story Point estimation plays a greater role.
Whether using capacity-based or velocity-based planning, these methods differ from traditional milestone-based planning as the team determines how much work can be done and not management. This is both a process a cultural change for most organisations adopting agile practices. When planning is done by management decree it is less likely that teams will deliver work to plan.
Short-term planning such as for a Sprint lends itself to suit capacity-based planning, as velocity tends to be variable and provide a less accurate forecast for a given Sprint. For longer-term planning, such as over several Sprints or several months, velocity-based planning is more suited as it relies more on average delivery rates (where work is less known). [4]
As always, commit to one approach then review regularly at retrospectives. If your method of planning is not providing desired results then consider deliberate change.