The Lean Startup — Overview and Personal Thoughts

I read “The Lean Startup” by Eric Ries few months ago. I strongly recommend you to read the book in order to fully get the concepts written here. This is one of the books that every engineer should read.


I read this book with not much experience about management, entrepreneurship, startups or creating new products. However, after reading it, you can feel a bit more prepared to deal and face situations related to those topics.

When you start reading it, the concepts in your mind change considerably. I guess the first thing that captures your attention is the meaning of Startup.

“A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty.”

My preconception before reading this book was that “a startup was a kind of small ‘business’ trying to innovate”. It’s true that innovate might imply that you work under conditions of extreme uncertainty but that’s not the whole point. A startup is a group of people (who can be part of a larger organisation or they have just set up a new small company) that want to create something (product or service) without knowing the right answer or how the market is going to react to it (conditions of extreme uncertainty).

In my personal development as an engineer, I like to use this approach when I work on my personal projects as well as when I work as an employee at Capital One.

Suddenly, you realise that this approach is a human-centred thinking that allows you to be more efficient and learn about your successes and failures/opportunities.

From my point of view, it’s really powerful because it’s based on a Build-Measure-Learn feedback loop: instead of making a lot of assumptions, you can make constant and often adjustments with a steering wheel called build-measure-learn. You can quickly build something; test it out; measure the success of it and learn (with specific metrics) from what you saw.


Learning is a essential unit of progress for startups. The questions would be: What to learn? How to learn? Is it useful? Here it is when the term Validated Learning is introduced.

Validated Learning is a rigorous method for demonstrating progress when one is surrounded by extreme uncertainty in which startups grow.

It’s a empirical method in which the team discovers valuable truths. It is more concrete, more accurate and faster than market forecasting or classical business planning. It is always demonstrated by positive improvements in the startup core metrics.


When it comes to experimentation, you need to define a clear hypothesis that makes predictions about what is supposed to happen. This hypothesis is often guided by the startup’s vision. Even if the experiment is a failure, they are useful: provide instructive feedback and can influence strategy.

The two more important assumptions entrepreneurs make are:

  • The Value Hypothesis. Test if your product or service actually delivers value to customers.
  • The Growth Hypothesis. Test how new customer will discover your product or service.
Since this approach is very customer focused, when you define your hypothesis you should ask yourself: Do customers realise/know/recognise that they have the problem that I am trying to solve?


We briefly mentioned the Feedback loop before. Take a look at the image and think about it for a few seconds.

The Build-Measure-Learn feedback loop

What’s the entry point? There’s no entry point. So, where should I start? It doesn’t really matter! You might have a product or a prototype in which you work on it in the past and now you can see an opportunity in the market. Or maybe, you have the coolest idea ever so you can start building a product. Or you’ve learnt something from this data and can give you some ideas. The important thing here is that you should never get out of the loop.

The faster you iterate through the Build-Measure-Learn loop, the more you will learn and the more efficient your startup will be.

Minimum Viable Product (MVP)

Everyone has heard of MVP before but what is it?

The Minimum Viable Product “is that version of the product that enables a full turn of the build-measure-learn loop with a minimum amount of effort and the least amount of development time.”
What really struck me when I read about this was the fact that the first MVP of Dropbox was a video! Just a simple video explaining interactions and how it was going to work was enough to get funded. No dev work, no worries about tech difficulties… just a video! Left me speechless.

Learning milestones are an alternative to traditional product milestones and they are useful as a way of assessing progress in a accurate and objective way.


At some point, you have to define your Leap-of-faith assumptions. They are the riskiest elements of the startup plan.

Leap-of-faith assumptions are those kind of assumptions that you take for granted, you don’t know if they’re right or wrong, you just go with them.

The value and growth hypotheses are leap-of-faith assumptions and maybe they are the most important ones.

As a example, in the iPod business, one of those leaps of faith was that people would pay for music.


When you’re testing defining what you want to get from it is the most important part. Identifying your goals, what you want to learn and an early definition of success is crucial.

When building your MVP, remove any feature, process or effort that doesn’t contribute to the learning you seek. Constant feedback is a must-have since you can see how people react to the changes you’re continuously adding.

Early adopters provide a huge benefit to your product/service. They accept — in fact prefer, that’s why they are early adopters — that your solution is not perfect but they help you identifying those areas you’re missing.


You have to measure if you want to learn. You can rely on metrics for that, which kind of metrics?

Use Actionable Metrics. It must demonstrate clear cause and effect. If not, it’s a vanity metrics. In this way you can learn from your actions.

Avoid Vanity Metrics. They’re easily manipulated and do not necessarily correlate to the numbers that really matter.

Remember that your metrics should be:

  • Accessible: Everyone should understand the reports. Make those metrics accessible to everyone so they don’t have to spend a lot of time learning how to use that data.
  • Auditable: We must ensure that the data is credible to employees.

You can user a Cohort Analysis to represent the analytics. Instead of looking at cumulative totals or gross numbers such as total revenue and total number of customers, one looks at the performance of each group of customers that comes into contact with the product independently. Each group is called a cohort. Let’s take a look at one example:

Cohort analysis example

If you see the image, probably we don’t care about how many logins we’ve had in our website. Why? It might happen that we’re getting more early adopters! Imagine that we’ve added a new feature to our website, how does it affect the people who pay? With vanity metrics as number of logins we cannot know. If we study different cohorts we can see how a new feature impacts the product.


I’d like to give some personal opinion here. Sometimes we get a lot of different feedback from customers. How to deal with it? This is my preferred personal approach:

  1. Identify the problem and see its consequences. Is it worth fixing/solving/improving? If so…
  2. Establish a metric that allows you to know if that problem is solved or not.
  3. See different options and weight on how they improve/solve the problem and how it affects the previously defined metric.
  4. Implement the change.
  5. Measure the change. Was the goal achieved? If not, go back to 3.

Pivot or Preserve

Each cycle in the build-measure-learn loop should make progress. If we’re making progress and a considerable good progress we should keep what we’re doing (Preserve). You should ask this question every time you complete a cycle. What if our metrics tell us that we’re not progressing? Maybe it’s time to pivot.

Pivot is a correction designed to test a new fundamental hypothesis about the product, strategy and engine of growth.

When a company pivots, it starts the process all over again (reestablishing a new baseline and tuning the engine from there). The sign of a successful pivot is that the new experiments you run are overall more productive than the experiments before the pivot.

How does it work? Establish the baseline, tune the engine and make a decision to pivot or preserve.

A pivot requires that we keep one foot rooted in what we’ve learnt so far, while making a fundamental change in strategy in order to seek even greater validated learning. It unlocks new opportunities for further experimentation/investigations which produce new ideas (new hypothesis) to be tested.

Ask most entrepreneurs who have decided to pivot and they will tell you that they wish they had made the decision sooner.

Vanity metrics prevents pivoting since the learning got from them doesn’t show how the product is performing, they form false conclusions and they live in their own private reality.

Types of pivots:

  • Zoom-in pivot. What was considered a single feature in a product becomes the whole product.
  • Zoom-out pivot. What was considered the whole product becomes a single feature of a much larger product.
  • Customer segment pivot. We’re solving the right problem but for a different customer than we originally anticipated.
  • Customer need pivot. The target customer has a problem worth solving but not the one we originally thought.
  • Channel pivot. The same basic solution could be delivered through a different channel with greater effectiveness.


By identifying which engine of growth a startup is using, it can then direct energy where it will be most effective in growing the business. Three engines of growth:

  • Paid. Advertising. This is what most businesses rely on.
  • Viral. Word of mouth! Your product advertises itself. Either by telling their friends or simply using your product, your customers will do your advertising for you.
  • Sticky. You want to retain customers for the long term.

Small batches

The biggest advantage of working in small batches is that quality problems can be identified much sooner. The startup can minimise the expenditure of time, money and effort so they’re not wasted. You can discover the truth faster.

Large batches tend to grow over time which causes the so called Large-batch death spiral. Moving the batch forward often results in additional work, rework, delays, interruptions, …


An adaptive organisation should adjust its process and performance to current conditions. The primary changes required are in the mindset of its employees, changing the culture is not enough.

One way of identifying problems is with The Five Why’s technique.

At the root of every supposedly technical problem is a human problem. Five Why’s provides an opportunity to discover what that human problem might be.

You start with a question: “Why have the app crashed in the same place 50% of the time?” And you find answers to it, the answer becomes another question: “Because it is caused by a NullPointerException” “Why is it caused by a NullPointerException?” “Because the developer doesn’t have enough experience with the code.” “Why the developer doesn’t have enough experience with the code?” “Because no one explained the code to him and he’s just guessing” “Why did no one explain the code to him?” “Because we don’t have an induction program” “Why don’t we have an induction program” “Because our manager doesn’t consider it necessary”

Blame inevitably arises but if a mistake happens, shame on us for making it so easy to make that mistake.

The five why’s approach acts as a natural speed regulator. The more problems you have, the more you invest in solutions to those problems.

In the example, you can invest some money in that Induction program, if that is again the root cause of other problems, you invest more and more money till it’s good enough.


“The Lean Startup” is the first book I read about entrepreneurship. It really helps you to identify priorities, hypotheses, focus on what’s important, etc.

This is a brief overview, I do recommend reading it since it provides more examples and you can see how different decisions can impact your business.

Hope you can get an idea of the main concepts and it can help you within your company or your personal projects.

Thanks for reading,

Manuel Vicente Vivo