The Lean Startup
Intro
As always, I don’t agree with or use everything I read, but I will still try and summarize it here. If you are new here, read why I put my semi-personal book notes online.
This is a book almost every single one of my friends has read. Almost everyone in every startup/entrepreneur program I’ve been in has read this. It’s a ubiquitous book within the startup canon.
Summary
In extreme uncertainty, a startup is an organism meant to create a repeatable and scalable business model. The author presents the “The Lean Startup Methodology”.
Vision
A startup has a direction it is going, aligned with its vision. The strategy to reach that goal includes:
- business model
- product plan
- marketing strategy
- partners, competitors, customers
Pivots are the strategic turns of a startup.
Validated Learning
Because things are constantly changing, one must continually validate that something is true. Launch experiments quickly learning both what works and what doesn’t. Remove as quickly as you add, especially early on.
This means you should be talking to the customer all the time. I see this as being one of the most common mistakes made by everyone. Understanding your customers is easier with empathy, something mentioned in multiple books including Leaders Eat Last. Although the context in that book is regarding leadership, I think it can be applied to any circumstance when you are interacting with people.
Monetize Early
Early monetization is a form of economic validation. Even if the amount of money is low. Waiting a long time to charge for a product causes a diversion in what you are building vs what customers want. Customers will want anything, if it’s free. But you need to focus on the feature customers still want when they have to pay for the product.
The initial revenue target can be low to start but grow steadily from this point.
Experiment
Use the Zappos method. Instead of starting Zappos by buying 1000’s of shoes, take pictures of 1000’s of shoes and post them online. Then when someone buys, go to the store to buy the shoes to then ship to the customer. This is the type of mindset to be in.
Scientific Method
The experiments need to be built in the right way by following the steps; build, measure, learn. For a startup, this cycle needs to be running constantly.
Build
Entrepreneurs want to get the most developed product for release. After all, it’s their baby. Release whatever you can to get feedback. Sometimes it comes out of a fear the entrepreneur has. That’s why one of my mantras is, “perceived risk is less than actual risk.”
Don’t be afraid of product managers from your competitors, they are likely too busy with their internal issues. And that’s assuming that the company even has a product group that has been told to prioritize this type of discovery. Most of the time, that company would rather just acquire you anyways.
Groupon
Groupon pivoted from digital activism to a manual coupon blog. Only after this coupon blog was tested did they then build automation. The blog allowed them to release and validate the market.
Dropbox
A slightly different story where Drew was raising money on the assumption that the cloud storage market was bad due to the customer experience. To get investment, he showed a video simulating the new approach to file storage. This allowed him to validate there was a demand for the product he wanted to build before actually going to build the software.
Measure
Test the riskiest assumptions first. If you are an engineer, that’s probably the market side. If you are an engineer, take it from me, you need to de-risk the marketing side. One type of de-risking not often talked about is how to de-risk the engineering side. And that’s because I think most current products are really about leveraging existing technology for the business side. The engineering risk is relatively small. However, many recent teams I’ve lead have a machine learning component to them which carries a huge technical risk. Check out how to implement a method of engineering de-risking called “tracer bullets” The Pragmatic Programmer.
To measure, you need to pick a fundamental metric. This is very similar to machine learning. This metric should express the real value of the product to your customer such as new customer acquisition or revenue increase. Good metrics should be:
- Actionable
- Accessible
- Trustworthy
Forget vanity metrics such as registered users.
The author recommends creating a template to prioritize your assumptions grouped into multiple categories. Personally, I use pending tests, in-progress tests, and validated tests. Trello is a handy tool to track this.
Pivot
Track your run rate closely, this tells you how much time you have left. But how do you know when to pivot? If things are going well, your important metrics are growing, keep going. If you aren’t making progress within the vision you have set, it’s probably time to pivot in a new direction. If you are new to this space, you should know that successful companies tend to pivot at least a few times. It’s not all that uncommon.
There are many different types of pivots, some of which are very relevant for existing companies with larger teams.
Treat the pivot like an experiment. Before you pivot, de-risk by creating an experiment.
Zoom-in Pivot
When one feature in your product is so useful you can pivot by making a new product with only this feature.
Zoom-out Pivot
New features are integrated into an existing product to create a new product.
External Pivot
The current MVP becomes a feature in a product
Customer Segment Pivot
The existing product is marketed to other customers. Pricing and customer channels should be reviewed. Hawthorne wipes if you are familiar with the community would be an example of this.
Customer Need Pivot
The customer may be the same but you solve a different set of problems for that customer.
Platform Pivot
Change from a single product to a platform for other products. I think an EHR would be a good example of this.
Business Architecture Pivot
From Crossing the Chasm you might switch from high-margin low volume to low-margin high volume. According to him, you can only be one but not both.
Value Capture Pivot
How you make money may change. From a subscription service to becoming ad-driven is an example of this.
Growth Style Pivot
When you change from one growth style to another. Common options are:
- Viral - current users recommend other users
- Paid - use marketing to acquire new customers directly
- Sticky - low churn rate means you don’t lose customers
Channel Pivot
Changes to how and where you sell your product(s).
Technology Pivot
Use new technology to achieve the same outcome. Beneficial if the new solution has some sort of value benefits like increased speed or decreased maintenance cost.
Product Market Fit
If you are in doubt whether your company has product-market fit, you probably don’t have it. Once you have market fit it’s time to find new channels of acquisition and increase the rate as which you improve the product.
Outro
I can already tell I’ll be revisiting this book every so often. There is so much in this book, you’ll want to read it all if the above summary looks interesting or relevant to you. I do like that this startup focused book seems to have some very relevant advice to leaders of teams in bigger companies that want to innovate.
