Lean Startup

[prpl_presto_player _builder_version=”4.14.7″ _module_preset=”default” theme_builder_area=”post_content” video_id=”52173″ hover_enabled=”0″ sticky_enabled=”0″][/prpl_presto_player]


Lean Startup is an iterative learning approach to building a scalable new business, combining Agile and Customer Development practices. The term was coined and trademarked by Eric Ries in 2008. Its principles were defined on Ries’ blog, Startup Lessons Learned, and in his wildly popular bestseller, The Lean Startup. These sources, however, didn’t really teach how to actually do lean startup. It was never really a “methodology”. It is a framework.

The best source for how to do it was actually found in a Google Group, where a bunch of startup people shared their best lean startup practices.

Here are the core principles as defined by Ries, with my take at describing what they mean:

Ries' Principle

My Comment

Entrepreneurs are everywhere

Everyone is an entrepreneur, provided they face uncertainty.

Entrepreneurship is management

Peter Drucker pointed out entrepreneurship is a way of organizing and managing a business in Innovation and Entrepreneurship. Organizing, managing, and performing work differ when faced with uncertainty.

Validated learning

The components that make up a successful business can be tested and verified.

Build, measure, learn

An iterative feedback loop is used to develop products, meaning teams build part of a product, measure whether customers care, and learn what needs to be changed or improved, if anything.

Innovation accounting

Progress isn't measured by tasks performed over time, but rather progress is being made toward expected outcomes.

Ries trademarked the term, with the hope that “lean startup” didn’t simply become another buzzword without meaning. Hilarity ensued.

My favorite early criticism of lean startup involved defining each word in turn and using that to re-define what Ries meant by the term. “Lean” means “skinny.” “Startup” means “new business.” A “Lean Startup” is then a “skinny new business”.

But is that right o useful? Among many lean startup to this day means a company that doesn’t have any money. With such false logic, you can make anything seem absurd. For instance: “kinder” is German for “children;” “garten” is German for “garden.” Therfore “kindergarten” is a field where children are planted. (Barbarians!)

“Lean” can also mean not having much fat. To most athletes, fat is wasteful weight. A startup, at least in Silicon Valley, means to launch a new, scalable business. Lean Startup can then also be defined as “launching a new, scalable business without much fat.” Now we’re getting somewhere.

Why “Lean”?

In Ries’ original blog post,  “Startup Lessons Learned,” he refers to “Lean thinking” in the connotation of the Toyota Production System, or the way Toyota built cars in Japan starting in the 1950s.

The fundamental concept of Lean is to reduce waste. From this simple impulse, Lean Manufacturing evolved into a bunch of processes for reducing waste in the manufacturing of cars. They strove to use natural resources more efficiently, organize the assembly line for greater productivity, modify the product itself to reduce failures, and so on.

The idea wasn’t just about assembly line uptime, which might seem like the most obvious course. If you concentrate solely on the waste of downtime, you may unwittingly produce other waste that has long-term negative consequences. It might lead, for instance, to people making poor decisions that mask problems. They perhaps patch machine problems, rather than fixing them. They create workarounds to keep the line running, resulting in an increase in injuries. Or when the system finally fails, it fails massively.

Toyota took a different approach. Its idea was to empower employees to stop the line if they experienced a problem. If you solved problems as they were encountered, rather than patched them, the longer-term throughput would increase. Management looked to line workers to improve processes to increase efficiencies. Employees changed tool placement, managed materials flow, and modified work processes. They knew best since they were the ones who encountered the problems day in and day out.

The goal was zero waste. That could never be achieved, of course, but the idea of continuous improvement became the standard operating principle.

Ries’ original Lean Startup describes its characteristics as a startup that leverages open source and free software and cloud computing, applies the Agile development philosophy, and practices Customer Development.  As the cost of technology has dropped, we’ve seen a increased “democratization” of startups, where, quite literally, anyone can be a tech entrepreneur.

Agile and Customer Development are frameworks describing iterative learning principles as applied to software development, and finding the right market segment, respectively. As shown in the figure, customer development and Agile product development are two distinct but interrelated processes. Customer Development uses iterative learning methods that focus on testing and validating assumptions about who the right customers are, the needs they have, and how their needs might be addressed. Product Development uses iterative learning methods to ensure the product works and meets the customers’ expectations.

Generally, the product development process receives input from customers indirectly through Customer Development, and (when available) by measuring product use directly. The product development process iterates on the product continuously, releasing functionality directly to the customer as quickly as possible. The teams also run experiments to validate or invalidate assumptions about how customers will use the product.

The iterative learning process doesn’t stop at the product. The process is now used across the entire business model, not just for customer and product assumptions. The correct sales, marketing, pricing, distribution, and so on are also initially based on not-yet-validated assumptions and yet are critical to the success of the business, so must be tested.

Similarly, the iterative learning process doesn’t stop at a particular stage of the company. The process is used wherever there’s uncertainty. Therefore, it’s relevant at the earliest stages when one is considering what business opportunities to pursue through customer and product development stages, and continues to scaling, continuous improvement cycles, extending product end-of-life, and then again when looking for new growth opportunities.

You may no longer be a “Lean Startup” at this point, but you should continue to use the methodologies.

To summarize, a Lean Startup in practice:

  • Is trying to create a new scalable business
  • Faces uncertainty in doing so
  • Uses an iterative product development process
  • Interacts with customers in such a way as to discover and validate all aspects of a working business model

Build – Measure – Learn

Perhaps the single most important aspect Lean Startup brought into the innovation economy is the concept of purposeful experimentation. The rigorous application of the “scientific method” to the non-technical parts of a business, has contributed to the launching and scaling of numerous successful businesses, including Dropbox, HubSpot, Meetup, and Buffer.

As applied to startups and corporate innovation, the Lean Startup way of experimenting includes:

  • Formulating a hypothesis representing what you’re trying to learn: if we do x, y% of customers will do z
  • Leveraging specific experiment types, including landing page, “Wizard of Oz,” and concierge
  • Using experiment data to track early progress (innovation accounting)
  • Establishing the first version of a product as a high-fidelity experiment, known as the Minimum Viable Product
  • Defining the pivot, which refers to changing fundamental parts of the business model tied to learning

Rapid experiment isn’t merely rapid prototyping. It’s not “throwing stuff against the wall.” It’s not “seeing what works.” It is a rigorous process of measuring customer behavior to help indicate whether you’re on the right path with the product, pricing, marketing, and so on.

Ries’ background as a software engineer and so much of his writing focused on the technical side of a business, but the build-measure-learn. loop can be extrapolated to any idea realted to the business, where you must learn what will work in the real world.

  • Build an experiment (to test your idea)
  • Measure the results (customers experiencingthe experiment)
  • Learn if you got it right (extract lessons and insights from customer behavior)

What do you do next based on what you’ve learned?

  • Iterate (run another experiment on the same idea)
  • Pivot (change the idea fundamentally)
  • Press on (test the next idea)

By “idea,” I mean any idea you have for how your business will function that is not fully known and is necessary for the business to succeed. The idea is an assumption. You might assume that a group of active, middle-aged, professional women are willing to change their lifestyle over warnings from their doctors that they are pre-diabetic.

So you:

  1. Build a way to test the assumption: “Interview ten people who match the description, recommend to them a specific lifestyle change, provide them a “call to action” that demonstrates their willingness to change.”
  2. Measure the result: “Six out of ten say they are willing; two out of ten actually take the step to sign-up for a diet coach”
  3. Learn from the result: “Two are different from the other eight. Those two are similar in that someone in their immediate family suffered from diabetes.”
  4. Create a new idea: “A group of people who have a family history of diabetes will take a specific step to change their lifestyle to prevent diabetes.”
  5. Run the test again.

The iterative cycle from the Customer Development side integrates with the iterative build-measure-learn loop on the product development side. If you build an internet app, you can measure customer behavior online. That’s great data, but it doesn’t tell you why people act (or fail to act) in a specific way. You must interact with the user live to learn the why.

The same goes for complex solutions, or physical products. The same goes for other stakeholders inside your business ecosystem. In the diabetes example, that ecosystem might include patients, nurses, doctors, hospital administrators, insurance companies, regulatory officials, and so on.

The build-measure-learn loop should be applied to all critical assumptions across the business model. You want evidence before execution, regardless of whether you’re adding features to an existing mobile application or figuring out the supply chain for a crowdfunded hardware device.

Lean vs Lean Startup

It’s confusing that many people talk about Lean when they really mean Lean Startup or Lean Innovation. Generally, Lean is about reducing waste in known products for known customers. How one goes about being Lean in the known side of the universe is inherently different than how one goes about being Lean in the deep void of the unknown.

This discrepancy has caused much confusion in the market. There’s an entire separate (but related) industry dedicated to the former. (See lean.org).