Scaling Lean by Ash Maurya

The Book in Three Sentences

Early-stage products can’t be measured using traditional metrics like revenue and profit; a better metric is traction. Modelling traction enables you to recognise the one clear goal for your product. With your goal in mind, identify the constraints within your product (i.e. bottlenecks) and focus on improving the output of that constraint, repeating until you reach your goals.

Summary

Life’s too short to build something nobody wants.

Traditional measures of progress like revenue, profit, and return on investment (ROI), aren’t helpful during the early stages of product development for the following reasons:

1. Because revenue is near zero during the early stages, we settle for build velocity as a measure of progress. But measuring progress as execution of an untested plan is no better.
2. Investing heavily in quantitative metrics doesn’t automatically give you solutions. Metrics can tell you only what’s going wrong, not why. The more you invest in quantitative metrics, the more you end up drowning in a sea of nonactionable data.
3. Even when you are generating revenue, unless you can connect cause and effect, you can’t leverage the elements that are bringing you success, and you can easily be led down the wrong path.

A better metric is traction: the rate at which a business model captures monetizable value from its users.

You create value for your customers through your Unique Value Proposition, which is the intersection of your customers’ problems and your solution. The cost of delivering this value is described by your Cost Structure. Some of this value is then captured back through your Revenue Streams.

It is important that you run tests early in the business model validation process to ensure that you can also capture back some of this value as monetizable value that can be converted into revenue.

Customer Throughput is Traction

Throughput is the rate at which a system generates revenue through sales. In the case of a product, customer throughput is the rate at which nonpaying users are processed into paying customers.

Throughput Accounting

Let’s first more formally define each metric as it maps to the customer factory:

1. Throughput: the rate at which monetizable value is generated from your customers over their lifetime minus any totally variable costs such as the cost of raw materials—typically the cost of customer acquisition.
2. Inventory: represents all the money invested in the customer factory toward things it intends to sell. This includes things you expect, like your product, but also unfinished goods (users), finished goods (customers), equipment, and other infrastructure that goes into the manufacturing of these goods (e.g., servers, software, etc.). The term “inventory” is interchangeable with “investment” in your system.
3. Operating Expenses: the costs expended turning inventory into throughput. They include things like salaries and other costs incurred in the running of the system. The distinction between inventory and operating expenses may appear fuzzy. It helps to think of inventory as assets that contribute to the valuation of a company and everything else as an operating expense.

• Profit = Throughput – Operating Expenses
• ROI = (Total Throughput – Operating Expenses) / Inventory

The universal goal of every business is to increase throughput while minimizing inventory and operating expenses provided doing that doesn’t degrade throughput.

Traction Modelling

There are three stages of product growth:

1. Problem/Solution fit: Does your idea represent a problem worth solving?
2. Product/Market fit: Does your business model work on a small scale?
3. Scale: Can you realise the full potential of your business model?

The first stage (Problem/Solution Fit) is where you test for sufficient customer pull to get the factory started. The other two stages (Product/Market Fit and Scale) are simply stepped-up versions of the first stage.

Average time to go from idea to Problem/Solution Fit is eight weeks and around 3 years to get to scale.

The distance between each stage can be modelled using a 10x rule, which works both top down and bottom up. Essentially take your customer numbers at scale, divide by 10 to get numbers for product/market fit and again by 10 to get numbers from product/solution fit.

The Customer Factory Model

Before you can prioritize waste, you need to be able to see the factory floor. You can visualize the customer factory using five steps:

• Acquisition: turning unaware visitors into prospects.
• Activation: connecting your promise with the first user experience.
• Retention: The time’s a user returns to the product.
• Revenue: Where value is captured back from the user.
• Referral: Number of times a user shares to others.

The goal of a business is increasing throughput, which is not always the same as increasing customer throughput. Simply adding more customers will also drive up more inventory and operating expenses. You might be able to potentially drive up throughput by increasing prices instead.

Before applying an engine of growth to drive up your customer throughput, you usually have to first optimize the throughput levers in your customer factory. These levers are:

1. The batch size of users you allow into your customer factory.
2. The conversion rates of each of the five macro steps.
3. The time between conversions (or cycle time).
4. Your monetization (or pricing) model.

The model is sustainable when the monetizable value captured from your customers (LTV) exceeds the cost of creating a customer (COCA). For a healthy business, you should aim for a 3x ratio between LTV and COCA.

Measuring the Customer Factory

For a metric to be actionable, you additionally need to be able to derive causality. The gold standard for doing this is measuring your customer life cycle in batches (or cohorts).

Cohorts help you measure relative progress by pitting one batch of users against another.

You start by grouping your users into daily, weekly, and monthly batches based on their join date (or sign-up date). Then measure their significant user actions as they progress through your customer factory.

You can create cohorts by gender, acquisition traffic source, release dates, a feature they use, et cetera.

Identifying Bottlenecks & Constraints

At any given point in time, customer throughput is limited by a single
constraint.

Bottlenecks are generally where you find excess inventory (like users) piling up, examples of inventory pileups would be users waiting on you for something—support requests, promised follow-ups, et cetera.

Constraints can be broadly characterized as external or internal constraints. External constraints are market constraints. Internal constraints can be any of the following:

• Raw Materials: Unaware visitors entering the product.
• Time: The time it takes a user to move through the factory.
• Money: An accelerant, that lets you do more of what you’re doing.
• Equipment: Infrastructure needed to turn users into customers.
• People: Needed to build and run your product.
• Mindset: A constraint caused by a preexisting way of thinking.
• Measures: Different or conflicting goals and measures of success.
• Methods: Preexisting procedures and techniques for doing work.

Product Experimentation

A key concept from the scientific method is that guesses or theories can never be proven right— they can only be proven wrong. This is the concept of falsifiability.

Our job is to wield these guesses into strategies or growth hacks that make our business model work for some slice of time. The goal isn’t learning but achieving business results—aka traction. We do this by:

1. Using models to expose constraints.
2. Formulating ideas for breaking constraints.
3. Testing these ideas through experiments.

The real question isn’t whether you run experiments, but rather whether you run good experiments. This chapter will outline the ground rules for designing and running good experiments.

“If you simply plan on seeing what happens you will always succeed at seeing what happens because something is guaranteed to happen.” —ERIC RIES, THE LEAN STARTUP

You need to develop a culture that allows people to have strong opinions, strange hunches, and weird instincts that they can then rigorously test.

There is no such thing as a failed experiment, only experiments with unexpected outcomes. —BUCKMINSTER FULLER