After penning a paper on the insidious Sleeping Beauty problem last year, Giulio Katis returns to the Mule with this guest post exploring the central ideas of The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses by Eric Ries. Starting with the immediate application to business startups, Giulio develops to a broader view: dealing with uncertainty itself.
When you are about to undertake some activity, how often do you typically question what you are about to do?
If you are like me, typically you’ll just “Do it” (to quote one of Ben Stiller’s screen characters), but occasionally you’ll take the time to plan and reflect on how you can optimize what you are about to do.
We have been taught that when faced with an activity or a challenge we need to frame the problem, dissect the problem, plan a solution (if we are really clever, collaborate) and then implement. But what do we do when the problem is poorly understood, or if we can’t get the answers we need upfront? Pretend we know everything anyway? Give up? Have a stab and hope for the best?
In the context of doing business, Ries’ best-seller Lean Startup presents a systematic approach to dealing with this situation in business.
This book is part of a general trend to update traditional approaches to business management to accommodate the uncertainty and pace of change which new technology has created – which covers product, service and new capability development in most businesses today.
Ries sees the method he presents as a scientific approach to doing business that updates and complements the ideas presented in Frederick Taylor’s 1911 classic the Principles of Scientific Management (which championed the importance of analysis, planning and task specialization in business management, and influenced corporate legends like Henry Ford and Alfred Sloan).
Ries is by no means the first to do this. In any discussion on Startups today we take for granted concepts such as “disruptive technologies” and “disruptive innovation”, which have become part of our common language since Clayton Christensen wrote about them in his best-sellers The Innovator’s Dilemma and The Innovator’s Solution. And, as the name suggests, the practices of Lean Startup are to be understood in the context of the Lean approach to business process management, as pioneered by Toyota in manufacturing (which among other things challenged the assumption that optimal process engineering involved linear chains of specialised functions). Also, Lean Startup is closely related to software development practices such as Agile (with its iterative and disciplined approach to development involving continuous feedback and learning loops, as opposed to the Waterfall one-shot “gather requirements, design, build, deploy” approach) and Continuous Deployment (a process whereby all code that is written for an application is immediately deployed into production).
The basic ideas of Lean Startup, however, can be explained without references to these developed software and business management practices, and Ries does this in a simple, powerful and readable way.
Ries’ book is written primarily from the lens of a startup (which typically has to navigate extreme uncertainty with very limited resources); but as he makes clear the principles and methods are applicable to large enterprises, especially those that need to adapt to changing circumstances and operate in uncertainty in a cost-constrained manner.
Lean Startup comes from the perspective that the problem is not whether we can build or create a product, service or capability—we’ve become pretty good at building things that are well defined (perhaps part of the problem is that we’ve become so good at this); but rather the problem is what exactly should we build or create—which requires us to answer more deeply why we should create or support the things we are committed to, and question the assumptions that have been driving what we have been delivering to date.
So while many past business process management principles addressed the problem of how to optimally execute or produce and deliver a well-understood product or service, the problem Ries is solving is how a business operating in some degree of uncertainty can simultaneously explore, learn, build and service to maximise expected future value creation and/or growth in a resource constrained context.
The solution he presents deeply embeds the experimental method into the management process. In a nutshell, when developing, modifying or maintaining a product, service or capability, Lean Startup suggests we should proceed as follows:
- explicitly identify the assumptions driving the need (opinion is not fact)
- pick a key assumption yet to be validated
- create a set of metrics designed to validate and explore the assumption
- design a ‘Minimum Viable Product’ (MVP), which might be a change or an enhancement to the existing capability, that will allow us to obtain the desired metrics to test the assumption
- build and deploy the MVP
- collate the metrics
- review the validity of and re-consider the assumptions and what is being developed
This gives rise to the mantra ‘Build-Measure-Learn’ repeated throughout the book.
This feedback loop may sound like a recipe – but Ries points out that this framework is far from a recipe. Many of the steps above require critical thinking, context specific insight, brainstorming and in some cases courage.
On the point of courage, at the end of each loop there is a critical decision to be made which Ries describes in terms of having to choose whether to persevere or pivot. Pivoting involves “a course correction designed to test a new fundamental hypothesis about the product, strategy, and engine of growth”. Under Lean Startup, pivoting is not considered as failure (involving change of management, say), but rather a necessary and important part of doing business. Not pivoting enough before the startup (or project) capital runs out is typically the cause of failure.
This gives rise to the concept of startup time as opposed to calendar time. Ries notes that typically to measure how long a startup has left we take the capital left (e.g. $1mio) and divide by the burn rate ($100k per month) to get the answer (10 months); but an alternative measure, which may tell us something more about the likelihood of the startup’s success, would be to estimate how many Build-Measure-Learn loops or possible pivots the startup could perform before running out of capital. The central practical message of the book is that the faster a startup can get through a Build-Measure-Learn loop, the more it can learn and thus the greater the chance it will succeed before funding runs out.
What is learnt is obviously a function of both the questions asked as well as the way they are answered. In terms of the answers, a key distinction Ries draws is between what he calls vanity and actionable metrics. Vanity metrics (e.g. gross turnover, gross profit) are lagging indicators that tell businesses what they want to hear (until they don’t), and do not provide information that can be used to make constructive changes. Instead of focusing on these, Ries puts forward the concept of actionable metrics which are designed to answer questions about what is actually driving customer behaviour, turnover, cost, profitability etc. For example, actionable metrics on customer behaviour might give data on how the customer joined, what was their first experience, why they are leaving or being retained. As the name suggests, they provide insight into what needs to be changed to create more value and/or growth (and obviously should be used in any business, regardless of its size or maturity).
Perhaps one of the biggest challenges Ries’ asks (of anyone running a business) is to assess yourself not in terms of the quality of the products or services you have produced, and not even in terms of the growth or profitability you have achieved to date, but to assess yourself in terms of how much you have learnt about what is driving your customers, your costs, your profitability, your growth etc. To genuinely adopt this perspective would obviously require a radical and courageous mindshift for most managers.
How the Lean Startup method can be applied in a mature, large, complex business is not something Ries spends time on (Furr and Dyer’s The Innovator’s Method: Bringing the Lean Startup into Your Organization spends more time on this question). Even though this is a non-trivial problem, it would seem even in the context of a business unit that is focused on execution and optimization (as opposed to innovation), there is scope to apply Lean Startup methods. I say this because I believe there is a degree of uncertainty (and thus the need for learning) in just about all business areas. For example, in the NPR podcast From Harvard Economist To Casino CEO (which was brought to my attention by Mark Lauer quite some time ago), Gary Loveman describes his use of randomized experiments (e.g. A/B testing) in an established casino to understand what customers liked, what they didn’t, what would make them come back if they lost a lot of money one night, etc. (Gary Loveman was well-known, amongst other things, for recognizing the value of the repeat slot players over the high rollers.)
After reading Ries I found myself asking what the implications were for (business) strategy. It is often said that strategy is easy and implementation is the hard part. Nevertheless, there is still the myth of the business leader (read Steve Jobs) that had the strategic initiatives that guided the company exactly where it needed to go. But these types of strategic initiatives are typically just informed, inspired, or lucky guesses. If, however, a business leader can orchestrate the activities of their organisation so Lean Startup principles work concurrently along with all the other business management practices needed to effectively run their organization, in theory the strategic initiatives should evolve, accumulate, be generated by and selected for as a result of the way that the organisation operates and does business (read Build-Measure-Learn loops); with bottom-up (generative) and top-down (guiding, co-ordinating) forces connected by their own feedback loops.
Ries’ book is considered by many as a must read for anyone wanting to start up a business (making a couple of the Forbes top entrepreneur and business book lists in 2014); and no doubt will be on the reading lists (if it isn’t already) of many business managers in larger organizations that need to grapple with change and innovation. It’s also a good read for anyone who is interested in what’s going on “out there” at the moment in the land of entrepreneurs and business management theory. But I think part of the reason why it resonated so strongly for me (in addition to the practical value it has for my work) was that the book is written in such a simple and powerful way as to imply applicability and meaning more broadly than for business.
The importance of feedback and cycles in the Lean Startup approach should be obvious. Mathematicians, scientists, engineers and the military have long recognized the importance of feedback as a way of dealing with uncertainty (going back to Norbert Wiener, the originator of cybernetics). In fact, Ries mentions that the Build-Measure-Learn feedback loop owes a lot to ideas from manoeuvre warfare, in particular, John Boyd’s Observe-Orient-Decide-Act Loop. But even though these ideas have been explored formally for well over a century (and, no doubt, millennia informally), it feels like we have still a long way to go in understanding the role of cycles in nature. For instance, in 2011 the Edge asked a number of prominent thinkers to answer ‘what scientific concept would improve everybody’s cognitive toolkit’. Daniel Dennett’s response (which in my opinion was one of the most thoughtful responses the Edge received to the question) was the concept of cycles. As he ended his response: “a good rule of thumb, then, when confronting the apparent magic of the world of life and mind is: look for the cycles that are doing all the hard work”.
Fundamentally, Lean Startup is a study in how to deal with the unknown—both “known unknowns” through experimental design and measurement as well as (as much as is possible) “unknown unknowns”, through the process of continuous experimentation and exploration. In his 200 m.p.h. (and very readable) book Sapiens: A Brief History of Humankind, Yuval Harari asks the question ‘what potential did Europe develop in the early modern period that enabled it to dominate the late modern world?’. He makes the claim that (all the good arguments of Jared Diamond’s Guns, Germs and Steel notwithstanding) one way to understand Europe’s ability to expand and dominate was in terms of its approach to the unknown, as can be seen through the development of maps. He notes that before the fifteenth century unknown or unfamiliar areas were simply left out of maps, or filled with imaginary monsters and wonders. “These maps had no empty spaces… During the fifteenth and sixteenth centuries, Europeans began to draw world maps with lots of empty spaces—one indication of the development of the scientific mindset, as well as of the European imperial drive.” I would like to know (from someone familiar with this part of history) whether the European nations that were more successful at world domination were those that were in some sense able to more quickly and more effectively cycle through Build-Measure-Learn loops.
So, on reflection, the main message I took away from Lean Startup was not something specific to just business. Rather it was the reminder that no matter how much work we do to create certainty, the unknown is all around us—and that there are more and there are less constructive ways to engage with it.