The Lean Startup by Eric Ries

When the likes of Tim O’Reilly, Geoffrey Moore, Don Reniersten, Steve Bell, Dan Heath, Marc Andreesen and Mary Poppendieck all sing the praise of a book, it might be a good idea to check it out. Which #hypertextual did, bringing a review in the process. Eric Ries has written this book out of his intense experience with startups, with a strong focus on his experience as the CTO of imvu, an online application with 3D avatars.

Many people have struggled to make a start-up succeed. Few have come with such valuable and great insights. In a very Lean approach, Eric has built on all the events and failures that has happened during this experience to learn, learn and learn again. Out of this validated learning, he built a simple and actionable system for continuous innovation which he shares with us in this best-seller … Extreme Uncertainty

Ries has a a great definition for startups : organisations who develop a product under condition of extreme uncertainty. In this fascinating book, Ries accumulates many stories from companies of all sizes delivering different types of software products and shows how universal both this definition and the framework he proposes to govern it are.

Build, Measure, Learn and iterate

Ries is a big fan of the Innovator’s Dilemma and his framework applies the basic strategy Christensen recommends : build, measure (while testing the market) and learn to adjust. In order to save people time and money, it is critical for this loop to be as short as possible. As Christensen said, new markets can’t be analysed and no big plan can ensure for success. Ries has learnt it the hard way while building the first version of his system : 6 months of exhaustive work to build perfectly a system that nobody wanted. What’s the point in being on time and budget if nobody cares about your product ? As he states on p.90,  no amount of design can anticipate the many complexities of bringing a product to life in the real world.

This is the reason why his loop has removed the P for Planning of the very similar Plan-Do-Check-Act Deming circle : to avoid the risk of analysis paralysis. Instead, Ries recommends to launch very quickly the product and then measure how the customer responds using appropriate in order to measure so that you can learn and build the …

Validated Learning

Ries pitch is that a product aims to validate a hypothesis, what he calls the leap of faith. The data is validated thanks to measured customer feedback : this is validated learning. And the internet provides a fantastic laboratory to gather hard data as it allows to 1/potentially reaches millions of users 2/gather actual hard data 3/track different types of population 4/perform split-testing while deploying different versions of the product to limited users.

Even if that may not be as efficient as real users interview (the example Ries give on page 43 and the IMVU aha! moment is quite spectacular), these numbers are some kind of genshi genbutsu Lean principle, whereby the team learn from direct feedback from customers as it confronts to the brutal reality (a key of companies greatness according to Jim Collins). For this metrics to build validated learning Ries argues that they have to be actionable, accessible and auditable.

Nullis in Verba

In the 17th Century, Robert Boyle (often referred to as the father of modern chemistry) and Robert Hooke created the Invisible College. Their motto was Nullis In Verba (Believe Nothing From Mere Words). As Clay Shirky says in Cognitive Surplus they committed themselves to acquiring knowledge through experimental means and to subjecting one another’s findings to kind of scrutiny necessary to root out errors. They happen to be the precursor of the Royal Society. In Lean Startup, Ries somehow offers the Nullis In Verba for 21st century business strategy. The latter should be built using a small steps iterative and scientific approach based on empirical data which are the only inputs to take into account while deciding if the strategy should be deflected, or, as Ries says : should take a pivot.

The Pivot

Upon completing the Build Measure Learn loop, we confront the most difficult question any entrepreneur faces : whether to pivot or persevere

The measure step of the loop allows to evaluate how relevant and appropriate the hypothesis is, how valid the leap of faith happens to be. Based on empirical data, the pivot is the deflection to adjust the strategy. Pivot implement the nemawashi Lean principle whereby (as per Jeffrey Linker‘s book Toyota Way) the organisation make decisions slowly by consensus, thoroughly considering all options and implement decisions rapidly. Ries describes the different types of pivots : zoom in (narrow down the product scope), zoom out (extend it), Customer segment (change the target population) … Most importantly, Ries suggests another perspective on the startup life expectancy. Rather than asking How many money left do we have ? one should ask How many pivots left can we make ?

Continuous Deployment

A really interesting thing in this book is how agile approach aligns with this strategy.  Well, some of them at least. Ries states that at the beginning, it is not a critical issue to release a half-finished product. Remember that the first objective of this beta versions is not to deliver the finished article but to test an hypothesis. So quality issues are not so big at this stage (Ries is clear that they definitely are at a later one, though).

On the other hand what is important is the ability to deliver quickly. Not surprisingly the author mentions rapid development framework (such as Ruby on Rails), frameworks that allows to deliver web applications much faster (up to tenfold) than standard technologies. He also underlines the importance of a continuous deployment infrastructure allowing to release small batches of product updates very frequently : about a dozen daily releases at IMVU.

This fully automated process also happens to be the auto-immune system, identifiying any product errors, removing the change, notifying the corresponding team and preventing it from installing any new change on the system until the bug is fixed : this is the andon (stop the line) implementation of the Lean Startup. The beauty here is the perfect alignment throughout the whole organisation to foster Continuous Innovation : it starts from a Continuous Deployment infrastructure supporting a continuous flow. This also is Toyota principle according to Jeffrey Liker : create a continuous process flow to bring problems to the surface.

Continous Innovation

This is the end objective : set up an organisation applying the Build, Measure, Learn loop at scale to innovate on a permanent basis. Eric Ries shows in this book how it is possible to solve the main 21st organisation problem (as per Richard Florida in Rise of the Creative Class) : the tension between operational excellence and innovation. Ries recommends to set-up an engine of growth, build an adaptive organisation and lastly allows sandbox for new ideas to enter a fast BML loop without disrupting existing processes. A strongly recommended read. As an appetizer a video of a Ries talk at the Web 2.0 conference :


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s