Introduction: An Overview of Startup Strategy

A Brief History of Startup Theory

In 2005, Steve Blank, the godfather of modern startup theory, published The Four Steps to the Epiphany, in which he showed that a startup is not just a smaller version of a large company; it is a fundamentally different entity. While a company executes a known, working business model in an environment of predictable economics, a startup is a search for a repeatable, profitable, scalable business model under conditions of extreme uncertainty, where initial predictions crumble upon first contact with customers.

To navigate this search, an entrepreneur is to systematically and empirically validate each business model assumption of the Business Model Canvas (or the Lean Canvas, a variant specifically made for startups).

That same year, Steve Blank invested in a startup named IMVU on the condition that the founding team would sit in his Stanford class on Customer Development, the search process he formulated. After several years running IMVU as a Customer Development earlyvangelist, iterating on the body of startup strategy, co-founder Eric Ries published the groundbreaking The Lean Startup in 2011, merging Customer Development, Agile Development, and Lean Manufacturing, showing that the process we should use to validate business model hypotheses is the scientific method, with the downstream goal of minimizing waste.

Eric’s Build-Measure-Learn feedback loop involves creating a testable hypothesis of one of the business model assumptions, creating an experiment (e.g. an MVP) to test it empirically, and measuring customer response through a novel process called innovation accounting (which mainly consists of AB tests and cohort analysis), then repeat this loop with the next business model assumption if the experiment’s results matched the hypothesized predictions (validated by improvements in the startup’s core metrics) or iterate/pivot the business model assumption until they do.

Then, in 2012 and most recently in 2023, Ash Maurya extended Eric’s ideas and showed the ideal order of validating business model assumptions, along with a tactical process for doing so, in his 3rd edition of Running Lean, which is hands down the greatest text on startup theory ever written.

A Broad Overview

The ideal order of validating business model assumptions is determined by the chronological order of the business’s funnel steps, i.e. the sequence of events your customers will go through when they interact with your startup. While this is different for every business (and later on you‘ll want to methodically determine which specific events should make up your funnel), one common funnel representation developed by Dave McClure for SaaS startups called the AARRR Funnel contains elements most businesses share.

The order of business model assumptions to test follow the order of this funnel, starting with acquisition. Only once customers are making it to the next step of the funnel should you build experiments for that step, with the intent of minimizing waste and optimizing speed. One version of this for a SaaS product could be the following:

1. 1) Acqusition — Pick a channel for acquiring customers (e.g. ads, social content, etc) and ensure you can bring a flow of customers to your site, even if the landing page is a terse coming soon / waiting list page!
2. 2) Activation — Once people are flowing through to the site, add copy and a signup form to move people to a free trial. After they’ve signed up, all they see is a “We’re still building it and will let you know when we’re ready!” page. If people aren’t signing up, you iterate until they do, often with learnings from customer conversations.
3. 3) Retention — Only once people have started signing up to use the product do you then start building the actual product.
4. 4) Revenue — Once people have started using the product do you start charging for it (or you can have revenue before retention and test pricing first instead, depending on which you find riskier)

Most first-time founders will first build a product (step #3: retention) and then drive traffic (step #1: acquisition) and get people to sign up (step #2: activation). The risk is what if you can’t get people to sign up? Then it doesn’t matter that you built a product, no one is making it far enough along the funnel to use it! This usually requires iterations of early steps of the funnel, then inevitably iterations of the actual product, leading to waste. Since startups fundamentally die because of an exhaustion of resources, the goal is to validate business model assumptions with minimal waste. Therefore, always build in the order of the funnel.

A few notes:

1. 1) Baked into building experiments to move users from early funnel steps to later ones are business model assumptions you want to validate, e.g. channels, customer segments, customer problems, solutions, revenue streams etc.
2. 2) It’s fine (and encouraged even, for optimal learning) if each funnel step makes it seem like the next step of the funnel exists even if it doesn’t; if customers are upset for having been duped and you lose that customer, it’s okay because there are always more potential customers. If there aren’t, you didn’t have a business anyway!
3. 3) You should always have new users entering the first step of the funnel while you’re building out later steps, since stopping and re-starting the flywheel tends to lead to waste.
4. 4) To have a flow of customers reaching a later step, you either have to increase the volume of customers getting to the previous step, or increase the conversion rate from one step to the next (which includes decreasing cycle times). To accomplish this, you run Lean Startup experiments.

Once you’ve been able to consistently move unaware visitors all the way to the end of the funnel (usually revenue), you have reached Repeatability, a major milestone that we can call the end of Phase 1.

From here on out, all the way to being a billion-dollar business, there are fundamentally only 3 activities you can perform:

1. 1) Increasing the volume of people entering the funnel
2. 2) Improving the conversion rates of steps along the funnel
3. 3) Increasing Customer Lifetime Value (CLV) at the end of the funnel by providing new sources of value where value-received >> price-paid > cost-to-deliver

These activities will be executed by doing nothing but the scientific method loop described in Part 2 of The Lean Startup, activities I refer to as Lean Startup equilibrium activities.

The next milestone (end of phase 2) is profitability (or virality if your primary engine of growth is referral) so that you can reinvest the output of your funnel (dollars from revenue) back into the input (dollars for customer acqusition) and create a divergent feedback loop.

Phase 3 is scaling up and creating as divergent a feedback loop as possible, referred to as product-market fit. The resulting repeatable, profitable, scalable business model shows that all the business model assumptions have been validated, the search has concluded, and the startup ceases to exist, transforming into a company executing a known, working business model.

Now, let’s get tactical.