The Lean Startup
5 Key Concepts of The Lean Startup
An introductory guide to The Lean Startup business theory as the influential book is reaching its 10th anniversary
The Lean Startup by Eric Ries is approaching its 10th anniversary; published in September 2011, the book has been a tremendous success, selling over 1 million copies. The book has gave birth, or at the minimum popularized, terms, concepts and methodologies that have become the cornerstone of the day-to-day conducts of startups. Specifically, the term Minimum Viable Product (MVP), which is associated with Ries, has become one of the most popular terms governing discussions of innovation and early-stage execution. The charts below, taken from Google Trends, demonstrate both how the book is compared with two other classic books in the literature of innovation, as well as the popularity of the MVP term compared to other ‘competing’ terms.
As can be seen, even nearly a decade after, The Lean Startup maintain its ‘hall of fame’ position:
Google Trends analysis comparing 3 books: The Lean Startup / Ries (blue), Crossing The Chasm / Moore (red) and The Innovators’ Dilemma / Christensen (yellow) over 2010–2020 time period.
Google Trends analysis comparing the term MVP (blue) with other competing terms: agile (red) and devops (yellow)
With its 10th anniversary on the horizon, it seemed like a good opportunity to revisit the book and to try and reflect how and to what extent the business theory has been used over the last decade.
I plan to cover it through a series of posts. This post will focus on introducing the key concepts of the book and the theory.
Introduction to the Lean Startup
The starting point of this book is with the assertion of the author, that the accumulative knowledge body that we have on management and management theories is ill-suited to serve startups and its in the root of the reason most of them fail. While startups may look similar to corporations from the outset, Ries argues, they are quite “a different beast”.
He then turns to define the term ‘startup’ as “a human institution designed to create a new product or service under conditions of extreme uncertainty”; this broad definition is deliberate in its attempt to include other kind of “constellations” that go beyond our conventional thinking of a startup, such as NGOs or “skunk work teams” within larger cooperations.
Ries attempts to deliver a new theory of entrepreneurial management, which is rooted in the history of the lean manufacturing movement brought to fame by the Toyota Production System, only that in Ries’ case the focus is on the intangible product which is software, rather than on the hardcore production of parts and goods.
A Startup is a human institution designed to create a new product or service under conditions of extreme uncertainty.
The focus of the book is on building a product, and on the framework to optimize it with the objective to reach a market-product fit and growth, but it is given in the context of a broader framework that includes elements of Vision and Strategy, as depicted in the triangle below.
The vision is the “north start” of the startup which rarely changes. The strategy is what you employ to reach the vision, including the business plan, product roadmap etc. A strategy may change (i.e. the pivot) when it fails to deliver a product that (even following many iterations) do not produce growth. The product, the focus of the book, is what undergoes the entrepreneurial management process in order to generate growth.
The Vision-Strategy-Product triangle may seem synonymous with Simon Sinek’s Golden Triangle of Why (Vision), How (Strategy) and What (Product). The main focus of the book, then, is on the ‘What’.
The Lean Startup is a broad theory of management and I will surely do it injustice by trying to condense it into a single medium post, but I do want to highlight, what I consider to be, the five most important points of the theory:
#1 — What Does A Startup Produce
Considering a theory of management which is modeled on a systematic approach to production, one should ask what it is that a startup produces.
Intuitively, we would want to answer that a software company produces software products. A financial standpoint might suggests that what it produces is value to investors; still a marketing mindset will claim that it produces value to customers.
Ries rejects all those answers. To his method, what a startup produces is validated learning. In other words, under conditions of extreme uncertainty it is unrealistically to focus on the traditional metrics of value and success. Instead, a startup must regards itself as a learning organization and the outcome of the learning is to validate its hypothesis about value and growth, as defined by its strategy.
#2 — A Pair Of Hypothesis
A product strategy must poses an hypothesis, namely the assumptions that the learning is looking to validate. According to Ries, there are two such hypothesis, a value hypothesis and a growth hypothesis.
The value hypothesis tests “whether a product or service really delivers value to customers once they are using it”. In other words, it is a test for the market-product fit, which is described by Andreessen as when “the market pulls product out of the startup”. Every strategy should try to answer why would users use the product in the first place, and one of the things The Lean Startup provides is a scientific method to answer that question.
The growth hypothesis tests “how new customers will discover a product or service”; that is, it focuses on the engines by which the product will pull customers in, whether that is through viral growth, paid marketing or otherwise. This hypothesis is just as important as the value one, for one, because it allows us to protect ourselves against the known chasm of going beyond your cohort of early adaptors.
#3 — The Build-Measure-Learn Loop
Given a basic concept of the Startup ‘product’’ (validated learning) and its validation criteria (hypothesis) , the theory turns to define how these can be achieved through the build-measure-learn loop. The build-measure-learn loop is probably the biggest innovation of the book, and its most used tool in practice.
While the implementation sequence of the loop is as its name may suggest, the way you plan a feedback loop is actually reversed:
Learn — what does the startup try to learn in a given iteration of the product, related to its value or growth hypothesis. For example, are users interested in a certain feature (value), will people pay for a certain service (value), will users invite their friends to join the product (growth) etc.
Measure — what is the data to support that learning and how to collect and measure it. Ries goes at length on how to go about measuring and how to keep away from what he calls ‘vanity metrics’; this practice is referred to as ‘innovation accounting’.
Build — what is the product or experiment the startup needs to build and launch in order to test the hypothesis. This is where Ries introduces the term ‘Minimum Viable Product’ (MVP), likely the most known topic associated with the theory. An MVP is defined as “ that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort”. So in other words, the focus of MVP is not about quality or usability, but solely about the learning objective. What doesn’t support the learning is waste and should be thrown out.
An MVP is defined as “ that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort”.
#4 — Innovation Accounting
Part of the build-measure-plan loop is indeed measuring, but in practice it is not a simple task and if done in the wrong way can yield to misleading results, or in Ries’s language, to ‘vanity metrics’. For example, simply measuring the total number of users an application has will hide critical data points, such as how much of that growth stems from the engines of early iterations as opposed to what is driven from new ones. In other words, how much of this growth can be actually attributed to new changes made in the product (out of a new BML loop). This is a dangerous opacity that could hinder a long-term view by creating blind spots to engines of growth that may in fact be dwindling down without us noticing it before it is already too late.
Instead, Ries suggests a method he calls ‘innovation accounting’ which he defines as ‘a quantitative approach that allows us to see whether our engine-tuning efforts are bearing fruits’.
At the core of ‘innovation accounting’ are two key principles:
- Identify what is your engine of growth: sticky, viral or paid (personally I think there are more than three, but hey … that was 2007); this will allow one to define the right metrics to track.
- Measure it using ‘cohort analysis’ which allows mapping the data with the iterations of the product and to clearly identify the causality chain between the experiments and their contribution (or lack of) to growth.
#5 — Pivot Or Persevere
Lastly, there are cases where engines stop growing even after significant number of product iterations (as a side-note, Ries suggests to define a startup runway by the # of iterations left at its disposal before it runs out of money).
When engines begin to run out of fuel, it might be the time to consider pivoting, that is, making a change to the company’s strategy. Part of the innovation accounting, described earlier, is the notion of building a ‘learning milestone’, a recurring meeting following every so and so iterations, where the team decides whether the metrics justify a pivot or only a sustaining effort to ‘tune up the engines’ (persevere).
Pivots don’t have to look like turning a bar check-in app into a photo sharing services or a game company into an organization messaging app; pivots could, and often are, more subtle (“changing course with one foot anchored to the ground”). In practice, Ries gives over 10 different types of pivot, like a decision to focus on one feature of the original app (‘zoom-in’) or targeting a different type of user (‘customer segment’) as a strategy to ‘leap over the chasm’ and reach beyond the early adaptors.
Once a pivot is taken, the strategy is changed and adapted, leading way to a new set of value and growth hypothesis, and to new cycles of Build-Measure-Learn iterations that further tune the engine and grow the company.
The Lean Startup is a comprehensive theory and one of the more profound studies I came across about life and death of startups; there are other topics in the book, that I didn’t have the time to cover, such as the use of small batches, the Lean Startup Kanban board, the principles of continuous development and others.
If you haven’t yet, you should definitely read the full book.
In the introduction to the book, Ries highlights the failure of startups, beginning with his own account of his first failure as an entrepreneur. One of the sections carry the title “Why Startup Fails”. It would have been comforting to think that following the success of the book and the emergence of the lean movement and with so much popularity to the CI/CD and MVP methodologies it helped foster, that the failure rate of startups would decrease.
Sadly, this is hardly the case, and why it is such will be the topic of my next post.