Concept of decision making African American guy with arrows and wires
ATD Blog

Don’t Fail Fast, Learn Fast

Wednesday, December 5, 2018

The moment we think we know exactly how a project should be executed, that we’ve done it 100 times before, and that we have planned for all of the potential obstacles, the universe is there to give us a reality check.

If you participate in designing, developing, implementing, or managing a learning product, these situations probably sound all-too-familiar to you:

  • We get surprised by a new method, tool, or technology and it wreaks havoc on our carefully estimated timeline.
  • We miss some obvious dependency that complicates our budget or requires new resourcing.
  • We didn’t identify a requirement from a key stakeholder up front, but it emerges during a mid-point check and blows up the project scope.

When faced with situations like these, we try our best to prevent them from happening again:

  • We build complex systems and toll gates to lock stakeholders into designs.
  • We create insane approval requirements, negotiating every detail up front and insisting on signatures from executives to push through changes.
  • We try to afford for derailments by building extensive contingencies into our timelines.
  • We implement pricing models that tack on fees for every customization, no matter how minor or simple

Despite these creative new processes, intricate systems, and complex rules, Murphy’s Law continues to strike. What can go wrong, will go wrong. Or, as a wise project manager once told me, “Plans are lies committed to paper.”

But there is a different way to approach these problems: accept that some amount of failure is inevitable. Let go of the frustration of chasing requirements, timelines, and stakeholders. Instead of trying to project manage potential problems before they arise, adopt a mindset that focuses on minimizing failure’s cost and impact.

Over the past few years, I’ve been an advocate for applying Lean UX methods and tools to the talent and development field. Cross-functional teams collaborating and iterating to build great consumer websites and applications aren’t all that different from the teams who create e-learning and other learning experiences and manage LMSs or social platforms.


Here's a set of principles and practices aggregated from across the Lean and Agile landscape. They have been documented in excellent books and blogs by product management and UX thought leaders like Eric Reiss, Jeff Gothelf, David Kelley, Christina Wodtk, and many more. Let’s take a look:

Start with the user. As serial entrepreneur Steve Blank says, “There are no facts inside the building, so get the hell outside!” Great products start with understanding our users—their wants, needs, and goals. Using tools like proto-personas to summarize what we think we know about our users is one way to represent them in our process.

Make your assumptions known. You know what they say about assuming! Once we have a lightweight set of user and technology requirements defined, we can hunt through them to find the things that might trip us up later in the process. Assumptions are guesses or unknowns, like the idea that a user prefers a specific modality or format for training, that a pre-defined duration is required, or that training is even the solution!


Experiment and learn. Assumptions exist on a spectrum: How much do we or don’t we know, and what are the consequences of being wrong? The assumptions that present the most risk are the ones we need to test first. A Lean approach advocates for using a 7th-grade science throwback: the hypothesis statement. “If we do this thing for this user, we’ll see this outcome.” For example, one hypothesis might be: “We believe that new employees at our company will be more engaged if we give them access to senior leaders during their onboarding process. We’ll know we’re right or wrong when we get positive comments from employees/learners during our mid-point survey and see long-term increase in employee retention.”

Creativity loves constraints. Using methods like Collaborative Design, we can engage cross-functional teams in a process of ideation, critique, iteration, and convergence to generate new and exciting solutions. If your stakeholders are involved in its inception, the solution will have greater adoption and be more likely to have the desired impact. Following the example above, we might ask:

  • What kind of access to leaders would improve engagement and retention? Video welcome messages? Presentations? Focus groups?
  • What commitment will the organization be able to bear? Face to face? Virtual? Monthly? Quarterly? On Demand?

(You can learn more about the Design Studio by checking out my article in TD Magazine,Welcome to the Design Studio.”)

Get feedback early and often. I often hear from leaders that a critical gap in our field is the lack of feedback loops in the design processes. Getting our work in front of users early and often is critical to designing systems, processes, or courses that not only meet their needs but achieve our ultimate goal: changing learner behavior. Before you launch your courses, run some usability tests!

I’m excited to be back at ATD TechKnowlege 2019 to talk about product management, collaborative design, and usability testing during my session, Build, Measure, Learn: Lean UX for Instructional Design. See you there!

About the Author

Becca Wilson is a product manager, designer, and facilitator who is passionate about helping digital teams transform the way they work. She has 10 years of experience in instructional design and developing blended learning strategies for Fortune 500 organizations. She currently works at IBM on the Priority Programs team, focused on addressing the scarcity of AI skills in the marketplace and accelerating the implementation of AI into business. Previously, Becca was an education product manager at General Assembly, supporting the ongoing discovery and development of scalable learning products in UX and product management.

Be the first to comment
Sign In to Post a Comment
Sorry! Something went wrong on our end. Please try again later.