The Lean Startup
Eric Ries
The five Principles of the Lean Startup method:
Entrepreneurs are everywhere
Entrepreneurship is management
Validated Learning
Build-Measure-Learn
Innovation Accounting
Part one: Vision
Chapter 1
Start
Entrepeneurship is really management. Many entrepreneurs are afraid to apply strict principles in the beginning of a venture, as they are afraid of stifling creativity. At the same time, too strict planning in the beginning of a startup is inefficient as a startup is too unpredictable. Both leads to failures of startups and our society wastes time and other resources.
Lean Startup is takes its thinking from Lean, Design Thinking, Agile and other methodologies.
In stead of measuring productivity in production of goods, Lean Startup uses validated learning as a measure of productivity.
Lean Startup teaches you to run a startup where instead of creating plans based on assumptions, the Build-Measure-Learn feedback loop steers you in the right direction, whilst having a vision.
Chapter 2
Define
Once your team is set up, The Lean Startup methodology helps you define what process to use and defining performance milestones.
Definition of a startup:
A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty.
The uncertainty is what makes traditional management techniques unsuitable for startups
Chapter 3
Learn
Avoid wasting time developing a product nobody wants by producing validated learnings in high frequency, early on.
Validated learnings act as guides for your business plans. These are findings that you wouldn’t get by setting up a business plan upfront the traditional way.
Validated learnings are measures of productivity in a startup, rather than looking purely at adoption, profits, etc.
Chapter 4
Experiment
The learnings of an experiment is a product:
Vision → Assumptions → Experiment ↕ Prove or Disprove (Fail) → Learn.
These insights and learnings can influence your business strategy, at the same time as it is being developed, if the small-scale experiments are launched early.
Not having a solid, set-in-stone plan is critical to a startup’s success.
(A startup can also be an initiative within a large, established company or institution. It is not necessarily a “garage” venture.)
Part two: Steer
Chapter 5
Leap
The most important hypotheses to test are the value and growth hypotheses. These are the ones that are most critical to the project’s success, the riskiest ones to get wrong.
The build – measure – learn cycle is the catalyst that the startup uses to produce a product. The minimum viable product – MVP, the version of the product that allows a full round of this cycle at minimal effort, can then be used to steer the strategy further.
There are no rules for which strategy will work, therefore we have to experiment.
The first challenge here is to find a way to test the product systematically, the second challenge is to do this cycle without losing sight of the startup’s vision.
It is also important to do field work to learn what customers want. From this we create a customer archetype. This archetype or persona is also a hypothesis, but it guides the decision making of what needs to be prioritised.
Chapter 6
Test
This chapter brings some good examples of success stories of companies that started with MVPs in different forms.
It can be the simplest website you can imagine. Or even humans doing the work of the product: a concierge MVP.
An MVP can also be a video demo describing the product, or even multiple MVP simultaneously solving end-user problems, and then the final one is developed from that.
It is tempting to do a high quality product to avoid bad news, but the danger is we build with our assumptions of what quality means to the end users. Quality might not even be important for them.
Common concerns about a MVP might be patent issues, getting the idea stolen and the MPV damaging the brand. Getting the idea stolen is unlikely and an option is to use an alternative brand to distance the MVP from the brand.
Chapter 7
Measure
Forecasts and milestones set the traditional way is not helpful for startups as they are too unpredictable. The challenge is to draw the right lessons from those milestones. Therefore we use innovation accounting with three learning milestones
MVP - data (where are we now? Establish baseline)
Tune the engine and attempt to move baseline :arrow_right: towards ideal
Decide to pivot or persevere?
“The myth of perserverence is dangerous”
If you decide to pivot, you start this whole process over again. But you can 'tune the engine' with just enough customers to get data from it (for example by spending only $5/day on google ads…)
What is really important is to do a cohort analysis instead of looking at all the data as a whole. Separate users into customer flows. This enables you to ask the right questions to these users when doing qualitative research and in turn make more productive experiments.
Beware of optimising instead of learning (because you might optimise the wrong thing) and vanity metrics (boosting the numbers or not being able to see why numbers improve.
Metrics should be AAA:
Actionable (they should show a clear cause and effect), auditable (should be possible to test wether data is consistent with reality - one way is to talk to customers to check) and accessible (metrics should be genuine and based on people and their actions - numbers are people too! Metrics should also be accessible to the company as a whole, not hidden away on file.)