Rating: 8 / 10
This is a fantastic book. I highly recommend The E-Myth Revisited as well.
As defined by Eric Ries, these are the five principles of the Lean Startup. They concisely summarize the main aspects of the method.
Startups have one destination: creating a thriving and world-changing business.
This is called the vision. To achieve it, we employ a strategy. The strategy includes a product road map, a point of view about partners and competitors, a business model, and ideas about whom the customer will be.
Your product can change often. Your strategy can change — that's called a pivot. But very rarely does vision change.
Everything you do in your startup is dependent on learning. You formulate an idea — usually based on what you've previously learned — and you test that idea. The idea represents your hypothesis, and the test is an experiment you do, which involves creating a product and measuring customer response.
In summary, these are the steps in the loop
When building a MVP, cut that which doesn't contribute to you learning what you want to learn. We want to go through the loop as quickly as we can, so we can learn as fast as we can.
And why do we do all of this? To learn whether we should pivot or persevere.
This summarizes the core parts of the book. Below are some sections that get into specifics; how to optimize the individual parts, how to accelerate the loop, et cetera.
In Lean thinking, value is anything that matters to the customer. Anything that doesn't matter to the customer is waste, so you should avoid doing it.
There's something called “pull” in lean production where, when you consume something from your inventory, a signal is sent to replace it. Then you don't have to keep a massive inventory; just what is necessary.
And “pull” connects to Toyota's just-in-time production method. “Each step is the line pulls the parts it needs from the previous step.”
This is how “pull” relates to the Lean Startup model: When we've made a hypothesis, our product development team should be engineered to design and run the experiment as fast as possible.
Customers don't know what they want, so we formulate a hypothesis about it, and we test it by building a MVP solution for it.
We want to learn exactly what our customers want. Anything else is waste and should be eliminated. This is called validated learning because it shows in positive improvements to the core metrics of your business.
It's also backed up by empirical data from real customers.
This is very important: test your assumptions. Do customers actually want what you think they want? Your business plan is based on assumptions. You want to test those as fast as possible. We start out with some assumptions, which are hypotheses to test. We build a MVP as quickly as possible.
An experiment is a product. It solves a real problem, and it has real customers. You don't just plan and do research, you act. You test it out.
The two most important assumptions are the value hypothesis and the growth hypothesis.
Split testing tends to work quite well for experimenting.
Innovation accounting in three steps
Innovation accounting turns your assumptions into financial models.
Reports on experiments should be actionable, accessible, and auditable.
Pivots are like course corrections that are designed to test a new fundamental hypothesis about the product, strategy, and engine of growth. The sign of a successful pivot is that your experiments are overall more productive than those you ran before.
When pivoting, we should remember what we've already learned. Yet, we must still make fundamental changes to our strategy to seek even greater learning.
There are many types of pivots.
Work in smaller batches also help prevent a large batch death spiral. Say you're working on the new release for a product. The longer you take, the greater you feel the expectations are, and the more you try to put in it... which just takes even longer! And the spiral starts.
Instead, work in small batches/single-batch flow. Do small, frequent releases.
Also, when you try to enforce large batch sizes and everyone just works for themselves and passes the result of their work to the next guy for further processing, it could cause huge issues as well. What if you made a lot of designs (which took a long time), and the engineer responsible for implementing them doesn't understand them? Then they'll have to either interrupt you, or redo the design. And if you're not available, they have no choice but to redo it. This is why so much work isn't created as it was designed.
Batch size is the amount of items that move between stages. A single-piece flow is where you have a batch size of 1. It's basically just where you process one item at a time.
Single-piece flow is faster than having larger batch sizes because we tend to underestimate the amount of time spent organizing, moving, sorting, stacking the larger batch. So single-batching would be faster for the entire system.
Single-batch flow is also superior because you have a finished product much sooner than if you do a larger batch size. Then you can check for errors, etc.
Working in small batches can reduce waste because you can check for errors faster.
Reducing batch size allows us to learn faster than the competition.
Where does growth come from?
Sustainable growth is achieved by your startup's engine of growth (not one-off stunts).
Ways to grow include
You can use the Five Why's to find root causes of problems. It's just about asking 'why' five times.
And you can make proportional investments based on how big the underlying issue is. If a minor bug was identified, maybe you don't need a week-long boot camp to educate your engineers.