Lean Six Sigma

Lean Six Sigma is a methodology to improve process in an entreprise. It combines the philosophy and tools of Lean and Six Sigma. Its main goal is to improve performance by removing operational waste.

Lean

Lean is a project management methodology focusing on identifying and minimising waste to maximise value. Originally for manufacturing, it has seen been adapted to software development.

Main ideas

• Is this step necessary? Does it add value for the customer?
• Think big, act small, fail fast; learn rapidly

Strengths

• Fast delivery loop
• Continual development

Weaknesses

• Weak statistical rigor
• Lack of robustess

Seven principles

1. Eliminate waste
2. Amplify learning
3. Decide as late as possible
4. Deliver as fast as possible
5. Empower the team
6. Build integrity in
7. Optimize the whole

Eliminate waste

Everything that doesn't add value to customers is considered as waste (muda).

• Partially work done: written code that has not been tested, integrated or deployed. The longer the wait, the higher the risk it become deprecated or non relevant.
• Overproduction: developing more features than needed, or that has never been asked, just "in-case" (feature creep, software bloat, etc.)
• Extra processes: unnecessary steps and meetings. Can include the write of mandatory documentation/reports, approvals from uninvolved third parties, etc.
• Delays: waiting for approvals or decisions. Can include downtime, when people or systems aren’t working.
• Task switching: assigning people to multiple projects at once. Switching context take time to refocus and increase cognitive load.
• Extra Processing: Doing more work than the customer values.
• Hand offs: passing the project from one team/individual to another force the new owner to re-learn the project. Especially true with poor written documentation. Increase the risk of losing or misunderstanding context, key information, etc.
• Defects: Errors requiring rework or scrapping. Bugs and other issues take more time to fix later than immediately.

Amplify learning

• Software development is a learning process: teams should emphasize on what they are learning and how fast.
• Do experimentations via quick prototypes (e.g. PoC) to learn by practice.
• Get user/customer feedback by doing short iterations.
• Build, measure and adjust in continual improvement loops.
• Team collaboration and mentoring are typically encouraged to learn from each other.
• Avoid long spec lists, upfront analysis and exhaustive bloating documentation.

Decide as late as possible

• Software engineering is uncertain and deal with complex systems, so projects are often unstable and unpredictable
• Making choice as late as possible let the team decide based on facts and experience rather than assumptions.
• This doesn't mean that planning should not be avoid. Iterative methodology helps to keep this decisional flexibility.
• Planification should include options and adapting to the current situation.

Deliver as fast as possible

• When version of a product is quickly deliver, concrete measures and feedbacks can be analysis.
• Learned mistakes can be fixed and new must-have features can be developed for the next iteration.

Empower the team

• Trust the experts and listen to them to help take relevant decisions.
• Avoid micro-managing and hierarchical decision-making (i.e. the manager decide, the worker execute).
• Keep the team involved and motivated

Build integrity in

Integrity define how the product/service is designed with quality and transparency in mind

Perceived integrity

• The system is usable, intuitive and coherent. It lies between the system and the user goals.
• Customer can have a transparent overview of the system: how it is advertised, deployed, accessed, priced, etc.
• Frequent and direct communication with users and customers help preserved perceived integrity

Conceptual integrity

• The architecture of the system is coherent, logical and well-defined.
• Its components are well-connected, so maintenance and extensions are easier.
• Components are balanced between flexibility, maintainability, efficiency, and responsiveness.
• Refactoring is essential to keep the system clean and simple.

Optimize the whole

• Software is not just the sum of its modules or components. It should be considered as a system on its own rather than isolated pieces.
• If parts of the product, software or system is treated in isolation (improve, change, etc), it can create frictions and bottlenecks with other parts of the process

Six Sigma

Six Sigma is a set of tool and techniques for process improvement, initially used in manufacturing. It focus on identifying and removing the causes of defects (errors) and minimizing variability (σ).

Defects and variability

Common cause

• Some process always include a slight variation due to the process itself
• This variation is previsible and stable in time
• It can be controlled and managed
• Example: due to the nature of machines and material, a coffee bag that should weight 250g can have a slight ±2g variations.

Special cause

• Some event create a great unusual variation, considering as a defect
• When it happens, we search the root of the problem and fix it instead of reworking the whole process
• Example: a sudden drop of client satisfaction is measured, just after the hire of a untrained employee

Frameworks

DMAIC

DMAIC is a framework for improvement cycle, used to optimize business processes and designs

• Define: describe the business problem, scope, objectives, timeline
• Measure: collect data, decide what should be measured and how
• Analyze: identify the root of the defect/error/problem, verify cause as effect
• Improve: build and test a solution to the problem (for example with PDCA methodology)
• Control: insure that the solution is stable and robust

DFSS / DMADV

Design for Six Sigma (DFSS) focus on the development of new products and processes.

• Define: describe the goals that are relevant for the customers and entreprise strategy
• Measure: identify goals, risks, feasability, timeline
• Analyze: understand development, design and architectures alternatives
• Design: focus and iterate on the most promising alternative
• Verify: test the design, set up pilot run and prototype, implement solution