As a term lean startup was first used in 2008. It was coined by Eric Ries on his blog Startup Lessons Learned, in a post titled The lean startup. Lean startup has its roots amongst high-tech startups in Silicon Valley and San Francisco, but has since been adopted by startups worldwide, and have also won great acclaim amongst large established companies and organizations such as General Electric, Telefonica, BBC and the United States Government, to name a few.
But before we get ahead of ourselves, let us start by defining the meaning of lean and startup. All too often we talk past each other without understanding the meaning of the words that we use. For example, have you ever stopped to ask yourself and your colleagues how they interpret the words lean, startup or innovation - a word often used but seldom understood.
Traditionally lean is associated with the Toyota Production System, Total Quality Control and Six Sigma, where the focus is on reducing waste and the number of errors that occur during manufacturing. However, in lean startup being lean means to use the least amount of resources whilst iterating towards a solution that addresses the customer's most pressing problems significantly better than competing alternatives. It is not about streamlining or incrementally improving the existing production system, it is about finding out what to build in the first place.
When it comes to the often misused word startup, the accepted definition in the startup community, comes from Stanford Professor Steve Blank, and reads as follows:
"A startup is a temporary organization designed to search for a scalable and repeatable business model."
What this means is that a startup does not know who will be the customer, what the final product or service will look like, how to price the offering, or how to best reach their customers. What a startup has are a bunch of assumptions.
While a startup is searching for a business model that works, an established company is fully occupied with executing on a finely tuned business model that figuratively speaking generates $2 for every $1 that's put in. Yes, we are talking about the golden cash cow that established companies tend to protect until their dying days.
The differences between searching for a business model and executing on a business model are as significant as night and day and affect everything from metrics, methods and tools to recruitment, incentives, leadership and definition of risk. And this is where lean startup comes into the picture. Lean startup is both a methodology (or set or methods) and a mindset that fundamentally differs from "business as usual".
The overall process is often illustrated in its most basic form as shown here (borrowed from Eric Ries book, The Lean Startup).
Lean startup is data- and experiments driven, where the aim is to turn assumptions into validated learnings as fast and efficiently as possible. Lean startup advocates a fail fast - fail often mindset in order to gradually learn and understand more about the pain points and constraints that the customer is experiencing.
The goal is to achieve a so-called problem/solution fit, i.e. we have something that the customer wants, and ultimately product/market fit, i.e. we have something that the customer is willing to pay for and where the cost of recruiting a new customer is significantly lower than what we can earn from that customer. Or put differently, to build a sustainable business, Customer Acquisition Costs have to be significantly lower than Customer Lifetime Value (CAC < CLV).
The key is to learn what works and what does not, as opposed to prematurely executing on the wrong business model. According to the data scientists at The Startup Genome Project (you'll find them at www.compass.co), the number one reason why startups fail is because they scale to early. Since scaling burns money like nothing else, scaling before reaching product/market fit, only results in a deeper hole in someone's pocket. While sales may very well increase, profitability will remain elusive.
Fundraising to promote something that the market does not want is a sure recipe for disaster. Boo.com, Webvan, Better Place and Pets.com, are only a few examples of failed startups that chose scaling to capture market share over providing customer value. They ended up burning hundreds of millions of dollars of other people's money in the process.
As always, the devil is in the details. Doing customer interviews in a way that introduces large amounts of biased data, or only relying on quantitative data while ignoring qualitative data, will only cloud judgement and increase the chance of failure. Since the most likely scenario is for the initial business idea or model not to rhyme with the market, lean startup practitioners do not spend valuable time writing 30-page business plans with five-year Excel projections to boot.
Instead, we quickly list our assumptions about the market on a business model canvas. Then we identify the riskiest assumption, that if it were not to hold true, would kill the business idea. Next, the focus shifts to turning the assumption to be tested into a measurable hypothesis. We design our experiment and set a target for what we hope to achieve. Then we "get out of the building" and run the experiment.
Based on what we learn, we update our business model canvas. All of which can be done within one week. Compare that to most large companies that take several weeks just to call a meeting in which the HiPPO (The Highest Paid Person's Opinion) often prevails over a data-driven approach.
If the assumption is not verified (which is the most likely scenario) we start from scratch with a new experiment - and keep at it until slowly but surely (or as long as the funding will take us) we crawl towards something that the market wants and is prepared to pay for.
The experiments are often referred to as MVPs (Minimum Viable Products) or MVEs (Minimum Viable Experiments). MVPs are used to test for user engagement and MVEs for initial user interest. Similarly, prototypes are used to answer the question "Can we build it", while pretotypes (coined by Alberto Savoia) are used to answer the question "Should we build it".
By putting a pretotype in front of customers, e.g. paper sketches, explainer videos or landing pages that communicate the value proposition, we can test for and measure initial user interest. Since the aim is to use the minimum amount of resources possible to maximize learning about the most risky assumptions, the pecking order should be pretotypes and MVEs followed by prototypes and MVPs.
There are many who have helped to develop lean startup to where it is today. The Business Model Canvas, which has become a key element of lean startup, was first presented by Alexander Osterwalder and Yves Pigneur in their book Business Model Generation, published in 2010.
Ash Maurya introduced his own version of the Business Model Canvas, which he calls the Lean Canvas in his book Running Lean, published in 2012. Another Business Model Canvas variation is the so-called Javelin, also known as the Validation or Experiment board, developed by Trevor Owens, founder of the Lean Startup Machine.
On the metrics side, Dave McClure, founder of the 500 Startups accelerator program has contributed to the ever growing lean startup toolbox with his Startup Metrics for Pirates, which Alistair Croll and Benjamin Yoskovitiz then elaborated on in their book Lean Analytics, published in 2013.
The above examples show that lean startup is a set of methods and tools under constant development, but with the distinction that they all share one common mindset, namely that of disciplined data-driven experimentation in direct contact with the customer to quickly and economically find a business model worth scaling.
Those who are the most creative in designing and running experiments to obtain validated learning, will be those who are the most likely to succeed. Remember, everything is an experiment.
Andy Cars, founder of Lean Ventures