Lean UX and the SAFe Program Increment Life Cycle
By Natalie Warnert
There’s a lot of confusion surrounding the definition of Lean User Experience (Lean UX) and how to incorporate both the Lean UX discipline and Lean UX experts into Scaled Agile methods like SAFe, while achieving fast feedback and a seamless user experience. When applying Lean UX processes at scale, feedback delays—which are far costlier than with smaller systems—magnify the confusion.
This article describes how to structure a Lean UX Center of Excellence (LUXCE) to establish design standard governance across the value stream. It also describes how to leverage cross-functional Agile teams to implement Lean UX as a self-organizing function of PI execution.
A Brief Overview of Lean UX
Lean UX design is a mindset, culture, and a process that embraces Lean-Agile methods. It implements functionality in minimum viable increments, determining success by measuring results against a pre-stated benefit hypothesis. That hypothesis is tested early in the process through collaborative design methods within the Agile team. The results are then fed into a minimum marketable feature (MMF), which is evaluated by real users, whose feedback is then incorporated into the next benefit hypothesis cycle. This continues refinement and learning throughout the Lean UX process (see Figure 1).Empirical feedback drives the decision to pivot or persevere in feature investment and centralized design standards; governance enhances consistency and autonomy of the Agile teams’ feature testing and implementation.
Implementing a Lean UX Center of Excellence (LUXCE)
SAFe advocates for Lean UX to be integrated at all levels of scaled development, from the value stream down to the Agile team. Cross-functional Agile teams are solution focused, versus centralized by function (coding, test, etc.). Each Agile team can define, build and test, and—when applicable—deploy elements of solution value.
One of the cornerstones of SAFe is to decentralize decision-making. In Lean UX, this means that most decisions should be made as close to the problem as possible. Therefore, implementing Lean UX functionality becomes the primary responsibility of each Agile team. While there are generally not enough Lean UX experts to collaborate individually with Agile teams, there should be enough to have strong relationships with a few teams and to be as close to embedded as possible in a hybrid, decentralized model.
A Lean UX discipline and expertise included in an Agile team is a vital ingredient for the implementation, collaborative UX design, and the decentralized decision making necessary in testing MMF hypotheses.
However, to address the potential of a lack of consistency and governance across the components of a solution, we suggest creating a small, centralized Lean UX Center of Excellence (LUXCE) for each value stream. Usually led by a lead UX expert who has influence across the value stream, the LUXCE collaborates to determine design standards (e.g., pattern libraries, user controls, navigation principles, corporate branding, style guides, and other governance), which are implemented as UX enablers across the value stream. By providing the Agile teams with the resources to support decentralized decision-making, they enable the new feature development vital to true team-level Lean UX implementation, while the LUXCE governs product consistency across multiple contributions.
Although each LUXCE operates within a value stream, there are additional responsibilities when scaling Lean UX across portfolios. Simply, the full portfolio solution set should behave as cohesively as possible. Each value stream should have its own LUXCE, informing Solution governance and consistency. But another layer of governance is necessary to align the organization’s portfolios. In these situations, organizations will have a portfolio-aligned Lean UX Community of Practice (CoP) to communicate branding standards and other portfolio-level considerations across all Solutions in collaboration with each value stream’s LUXCE.
Preparing for the PI
Before PI planning, the Lean UX thinking is incorporated into the economic prioritization of the candidate recommended to enter the PI. From the moment Epics enter the Portfolio Kanban funnel state, until they are implemented and done, the LUXCE is a valued partner of the Epic Owners and LPM. When a Lean Business Case investigates the outcome hypothesis of the Epic, a Lean UX expert from LUXCE elaborates on the user need, keeping the work outcome-driven and the MVP flexible.
Once an Epic is approved and reaches the backlog state in the Portfolio Kanban, LUXCE experts consult on further refining the Epic into associated features. The goal is to create MMFs whose benefit hypotheses are clear, measurable, testable, and able to fit within a PI. This elaboration provides an economic Weighted Short Job First (WSJF) measure to prioritize the Epic into the implementing Portfolio Kanban state. We can then prioritize its associated features against those of other approved Epics in the Program Backlog.
Before entering a PI Planning event (addressed in the IP iteration or earlier), it’s crucial to communicate the prioritized features to allow trains and teams to start planning how they will implement and test the hypotheses with customers in the simplest way. Leading up to the event, Set-based design and Principle #3 – assume variability, preserve options should be used to develop the benefit hypotheses to test early in the upcoming PI. The empirical data and feedback they yield will continually inform the collaborative design stories to realize the MMF. Transparency of the Program Kanban and a well-defined set of benefit hypotheses can be the difference between a successful and PI Planning and a disaster.
Just-in-time design artifacts are necessary to prepare for PI Planning, helping to explain the context and limitations the system is operating within. Their level of polish will look different with every team, train, and feature based on the familiarity with the problem. For example, it depends on the maturity of the teams and train but artifacts could include basic wireframes and rudimentary designs detailing user flows. This doesn’t imply the team should design the entire solution. Instead, it should do just enough to visualize ways to view a benefit hypothesis and understand how to test it within the technical and user constraints. This is the beginning of the collaborative design process, which will further inform the MMF to be developed and evaluated. This also provides an opportunity to identify and challenge/prune features where the goals or hypothetical benefits are not sufficiently understood.
Like the Architectural Runway, the Lean UX discipline, informed by the LUXCE, has a UX Runway of sorts. To advance quickly, the benefit hypotheses and MMFs to be tested need to be enabled. Lean UX enablers include building out design systems, libraries, and style guides and templates as enablers. The LUXCE provides these system guidelines for embedded UX experts and Agile teams to create emergent visual designs and simple wireframes that can be quickly transformed into working prototypes to test with customers. If these Lean UX enablers are ignored, the runway will be consumed and velocity to implement meaningful feedback loops will be stunted. For Lean UX experts to keep the system sprinting without overrunning the runway or creating a feedback bottleneck, it’s vital that capacity is planned in for UX enablers within Agile teams.
Participation in PI Planning
After helping to prepare for a successful PI Planning event, the LUXCE has another important role to play during the event itself. It needs to be fully present to communicate, consult, refine, and plan around the PI vision with the Agile teams and program and business stakeholders. It needs to coordinate the prioritized features, benefit hypotheses, and minimum necessary artifacts. Leadership and product management set the program vision and business context; systems architecture sets the architecture vision. The LUXCE also needs to communicate the UX vision and explain how it complements the strategies. For the design sets to merge with the technical constraints, and the business outcomes (and so the user needs are not forgotten) it’s critical that trains understand how the business, technical, and user visions fit together. Vision setting should be done in the morning section of PI Planning on day one. During team breakouts, the Lean UX experts should be available to consult the teams they work with. Design artifacts should have already been communicated and shared. But Lean UX experts need to facilitate the shared understanding, leading Agile teams toward a collaborative design process where they are the embedded experts. A Lean UX expert’s primary responsibilities include:
- Helping to break out stories from features
- Planning benefit hypothesis tests
- Identifying associated dependencies and Lean UX enablers
- Maximizing feedback incorporation loops into early PI iterations
LUXCE experts also play an important role in crafting the outputs of PI Planning with the Agile teams and trains. PI Objectives are outcome-focused, as are the end-user goals of the benefit hypotheses. To highlight program-level dependencies, including UX enablers, LUXCE must have a place for dependencies on the Program Board. Lean UX experts can also address this while embedded in their associated Agile teams. Their presence and contribution to PI Planning is vital to the train successfully achieving alignment on its shared vision.
Participating in PI Execution
The LUXCE continues serving Agile teams and the trains in PI execution. While serving as key stakeholders in the LUXCE, the individual Lean UX experts also act as embedded consultants within Agile teams. There’s a delicate balance between doing the investigative work to plan for upcoming PIs and the just-in-time work relating to current iteration development. Shared backlogs need to reflect it all while not calling out “staggered iteration,” as Lean UX experts can be working 1+ iterations ahead of development while supporting just-in-time team needs.
Staggered iterations and separate backlogs create a mini-waterfall and more handoffs than optimal or necessary. This doesn’t mean there isn’t more UX work to do ahead of development. But SAFe incorporates that work into the normal iteration cadence for increased transparency, and collaborative design with the Agile team and the LUXCE experts. In keeping with set-based design, options will be eliminated in the first few iterations, Learning and feedback from benefit hypotheses emerge through collaborative design and test execution. While weighing options and completing the stories to develop and test the benefit hypotheses in early PI iterations, later iterations can incorporate the feedback by building the MMF. During PI Planning, it’s important to allocate the capacity from the built-in empirical feedback to this endeavor so that the MMF—a compilation of collaborative designs and tested benefit hypotheses—can be released to test and observe in production on demand with real users.
When empirical insights inform MMF development near the end of the PI (integrating feedback at least once per iteration), Lean UX enablers come into play. A healthy UX Runway makes UX design a team sport and decreases Agile team reliance on date-driven deliverables. Requirements, such as pixel-perfect visual designs for every screen, can rapidly bottleneck development. Agile teams can make some of these changes themselves as the UX enablers’ maturity enable them to support established design standards more easily.
Other design deliverables that are deemed necessary (but that enablers can’t supplement) arrive just-in-time and only as needed. This minimizes waste while maximizing empowerment and decentralized decision-making. With capacity freed-up from not having to constantly deliver pixel-perfect designs, embedded Lean UX experts can now focus on new benefit hypotheses, feedback from previously released MMFs, and continually building out UX enablers. This work is then planned into the next PI and prioritized against the other work in the program and team backlogs.
The key to making the Lean UX discipline scale is that the LUXCE needs to collaborate with the teams and trains in all activities. Embedded Lean UX experts should be at iteration planning, backlog refinement, iteration review, and retrospectives. They need to be involved in estimating, too. By adding their perspectives and sharing their work early on and often with the Agile team, the common understanding and UX vision will assist in eliminating late discovery of constraints in function and usability. In conjunction, synchronizing cadence of feedback cycles and incorporating empirical user data into the MMF must be planned and executed throughout the PI, not only at the end. This feeds system demand and integrates frequent learning points with empirical evidence of a valuable, usable product.