Scaled agile framework

Development teams typically refine their backlog up to two to three iterations ahead, but in larger organizations the product marketing team needs to plan further ahead for their commitments to market and discussions with customers.[16] They will often work with a very high level, 12 to 18-month roadmap, then plan collaboratively with the teams for three months of work.[17] The development teams will still get into detailed refinement 2-3 iterations ahead, only getting into detailed task plans for the next iteration.[18]

While development teams have a number of frameworks that define how they should be agile, there is very little that describes this for management. SAFe delivers many of the same principles, such as cross-functional teams, to the groups that handle the more abstract levels of responsibility and planning (product and portfolio).[19] SAFe has also been criticized for aggregating too many disparate practices.[20]

In Scrum, the product owner is expected to assume responsibility for the full product life-cycle, including the return on investment of development decisions, as well as performance in market. On large-scale developments, the organization wants a view across multiple team backlogs, such as provided by a product manager.[21] Although SAFe assumes the product owner role sits with product management, it has nonetheless been criticized for separating product owners into the development organization.[22]

Agile frameworks are designed to enable the development team to be autonomous and free to design how they work. SAFe acknowledges that, at the scale of many tens or hundreds of development teams, it becomes increasingly chaotic for teams to fully self-organize.[23] It therefore puts some constraints on this, so that where teams are working on the same product, their deliverables can be better synchronized for releasing together, although this has been one area in which SAFe has been criticized.[21][22]

The SAFe planning cycle recommends including an additional iteration after a release, so that teams can improve their practices and are ready for the next planning increment. Earlier editions of SAFe also designed this to be a hardening iteration, that is to stabilize or harden the product before releasing it. This was predicated on the complications of working with large integration environments where dependencies meant that you could not test everything until the very end. SAFe was criticized for this as it represented an anti-agile or waterfall element.[24] This is not included in recent editions of SAFe.

According to its authors, SAFe is based upon ten underlying concepts, which are derived from existing lean and agile principles, as well as observation:[25]

1. Take an economic view
2. Apply systems thinking
3. Assume variability; preserve options
4. Build incrementally with fast integrated learning cycles
5. Base milestones on objective evaluation of working systems
6. Visualize and limit work-in-progress, reduce batch sizes, and manage queue lengths
7. Apply cadence (timing), synchronize with cross-domain planning
8. Unlock the intrinsic motivation of knowledge workers
9. Decentralize decision-making
10. Organize around value

In SAFe version 5.0, there are four configurations: essential, portfolio, large solution and full:[26]

• Essential SAFe is the most basic configuration. It describes the most critical elements needed and is intended to provide the majority of the framework's benefits. It includes the team and program level (which it calls agile release trains or ARTs).
• Large Solution SAFe allows for coordination and synchronization across multiple programs, but without the portfolio considerations. In earlier versions of SAFe, this level was referred to as value stream.
• Portfolio SAFe includes concerns for strategic direction, investment funding, and lean governance.
• Full SAFe combines the other three levels.

Scaled Agile provides certifications that cover different areas and knowledge levels.[27]