Why Experimentation Should Be Required for Initiating Projects
In December 2012 during The Lean Startup Conference, Beth Comstock, senior vice president and chief marketing officer at General Electric (GE), participated in the following video conversation with Eric Ries. GE is an example of how a large organization can implement Lean Startup for projects that demand innovative solutions, as explained in this blog post.
During the conversation, Comstock mentioned that what happened at GE is the realization of what Ries suggested when he said that “entrepreneurs are everywhere.” Essentially, employees in organizations can all be entrepreneurs.
Comstock pointed how the concept of minimum viable product is relevant to GE in the way the company innovates. For Comstock, innovation starts by understanding the customer and advising products accordingly. Creating minimum viable product served to speed up the feedback cycle. More importantly, it helped to balance the usual perfectionist engineering approach with the need to quickly show products to the customer.
From this conversation, I believe that experimentation should be required for initiating projects—no matter the organization’s size—for the simple reason that product features will more than likely be discovered incrementally and iteratively during the project's lifespan. This is contrary to what we do during project budgeting, which is based on a large work batch. Sticking to an upfront budgeting mindset while ignoring experimentation produces inferior products at best.
Speaking about the state of project budgeting, Tom Gilb said in his Lean Startup blog post, "If history is any guide, we are not going to change our irresponsible software investment culture in the short term."
To have an experimental mindset, organizations should treat project requirements as hypothesesinstead of baselines for budgeting and estimation. Product delivery must be quickly performed in order to obtain customer feedback. This feedback improves the hypothesis by discovering more about the product. In other words, you must treat requirements as hypotheses and use agile product delivery.
I'd like to know your comments on what barriers organizations might have to implement experimentation approach in their projects.