AI skills
large-scale-scrum
Use when asked to scale scrum across many teams with LeSS (Large-Scale Scrum) — one Product Backlog, one Product Owner, and one Sprint shared across all teams, deliberately minimizing added roles and process — as distinct from scaled-agile-framework's more prescriptive, role-heavy approach to the same scaling problem.
LeSS, developed by Craig Larman and Bas Vodde, scales [[scrum]] to multiple teams working on one product by adding as little extra structure as possible: one Product Backlog, one Product Owner, and one Sprint shared across all teams, rather than introducing new roles or process layers to manage the scale. Its guiding principle — "more with less" — treats each additional role, artifact, or process as a cost to be justified, not a default response to growing team count.
The core structure
- One Product Backlog, one Product Owner — for the whole product, regardless of how many teams are building it, keeping a single coherent priority order rather than teams working against separately maintained backlogs.
- One Sprint — all teams sprint on the same cadence and produce one integrated product Increment at the end, not team-by-team increments reconciled later.
- Feature teams, not component teams — each team is cross-functional and capable of taking a full customer-facing feature end to end, rather than being organized around a technical layer or component; this is what makes a single shared backlog workable, since any team can in principle pull any backlog item.
- Sprint planning in two parts — Part One brings representatives from all teams together to select and clarify backlog items; Part Two lets each team plan their own sprint backlog and approach in detail.
- Overall Retrospective — in addition to each team's own retrospective, a cross-team retrospective looks at the whole multi-team system's process issues, not just each team's local ones.
LeSS vs. LeSS Huge
Standard LeSS covers up to roughly 8 teams (around 50 people). Past that scale, LeSS Huge adds Area Product Owners, each responsible for one requirement area with its own area backlog pulled from the overall Product Backlog — the minimum additional structure the method's authors judge necessary once a single Product Owner can no longer meaningfully manage the whole backlog personally, still deliberately smaller than most alternative scaling frameworks' role set.
LeSS vs. SAFe
Both scale Scrum-style delivery across many teams, but take opposite philosophies: LeSS deliberately minimizes new roles, artifacts, and process layers, keeping as much as possible identical to single-team Scrum; [[scaled-agile-framework]] (SAFe) is more prescriptive, defining additional roles (Release Train Engineer), events (PI Planning), and organizational layers (Program, Portfolio) by default. LeSS suits organizations willing to restructure toward genuine feature teams and a single backlog; SAFe suits organizations wanting a more complete, off-the-shelf structure without that same degree of reorganization.
Common pitfalls
- Keeping component teams and expecting LeSS's single-backlog model to work — if teams can't actually pull any backlog item because they're organized around a technical component rather than customer-facing features, the single shared backlog becomes a bottleneck rather than a coordination tool.
- Adding roles "just in case" before LeSS Huge's team count actually requires it — works against the method's core "more with less" principle; each added role should be justified by genuine need at the current scale, not adopted preemptively.
- Running Sprint Planning Part One as a status meeting rather than genuine cross-team backlog clarification — undermines the shared understanding the two-part planning structure is meant to build before teams plan their own work in Part Two.
- Skipping the Overall Retrospective — without it, systemic, cross-team process issues have no forum to surface in, even though each team's own retrospective is working fine locally.
Learn more
- [[scrum]] for the single-team framework LeSS scales with minimal added structure.
- [[scaled-agile-framework]] for the contrasting, more prescriptive scaling framework.
- [[scrum-of-scrums]] for a lighter-weight coordination pattern that fits fewer teams and doesn't restructure backlog ownership.