Getting groovy with “Lean Startup”

Last week I taught a class on Lean Startup at Republikken. In this post I’m going to introduce the two tools I think you need to get started with Lean Startup.

The two most important Lean Startup tools — MVPs and Cohorts

I think you only need two tools to get going with Lean Startup. They are 1) Testing assumptions with MVPs and 2) Cohort Analysis. The rest of the theory is good and helpful but just these two ideas should get you 80% of the value of Lean Startup.

The hardest part — getting into the “Lean” mindset

The recurring theme in the class was trying to wrap your brain around Lean Startup ideology. It is simple to say things like MVP, pivot and testing but not so simple to apply those principles to your business. And the intersection between theory and doing, what I call actionable ideas, is what interests me the most.

Lean Teaching the MVP — MVPS are for learning, not impressing

My main concern is boiling the concept of MVP down as much as possible into actionable ideas so startups can immediately start using them. The very first challenge is convincing the startup the point of the MVP isn’t to impress, but to learn. For example, CoursAvenue wants to build a great user interface so people can easily and quickly find good classes (for example, in yoga or cooking) near their homes. But one assumption we need to test is whether people will book their course online through CoursAvenue. In order to test that assumption, we need the simplest possible MVP that will allow us to learn about our demographic.

Once you decide to use MVPs to test your assumptions, there are five main MVP choices.

Customer interviews — Best for canvasing early ideas, Place ad and interview responders

Landing pages (smoke test), Kickstarter, high hurdle (cash)Offer product for sale before it existsUse mockup of product to test customer interest Models and prototypes (popup store, food carts)Low-investment shop to sell real productsTest dim sum taste before opening a full restaurant Concierge — manually do work before automationDo all the work by hand to see if solution solves customer problemFood on the Table — Wizard of Oz — create front end, but do back end manuallyCustomers think a computer is doing all the work but in reality a human is Zappos — pretend to be a fully-formed web app, but do backend by hand

Then the process becomes much simpler. 1) State your assumptions, as finely detailed as possible and 2) Devise a MVP that will allow you to test the assumption as quickly and cheaply as possible.

Cohort Analysis

From Eric Ries: “Cohort-based reports are the gold standard of learning metrics: they turn complex actions into people-based reports… 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.”

Here is an example to make it more concrete (and also answers a student question about whether this works offline).

Republikken is a coworking space with a cafe and desks for rent. In this example, their marketing funnel looks like:

Step 1 — People come to the cafe, to sit in the public space, work and have a coffee

Step 2 — People request a tour of the coworking area

Step 3 — People sign a contract with Republikken to rent a desk

The cohort chart looks at percentages, not totals. The numbers across the top refers to weeks — we broke out all the visitors into cohorts by week. The first week, before running any tests, 100% of visitors came to the cafe, 30% took a tour and 10% signed a contract.

In week 2, we tested giving people a beer if they took a tour. Our assumption was that a beer would convince more people to take a tour. We further assumed if someone took the tour they would be more likely to sign a contract. As the chart makes clear, the percentage of people taking a tour went way up, but the percentage of people signing a contract did not change. So we learned that our assumption was wrong: more people on the tour did not translate into an increase in sales. Therefore, we will not continue to offer a beer.

In week 3, we tested giving people a piece of paper with the benefits of Republikken membership at the end of the tour. Our assumption was that a clear statement of benefits would encourage them to sign a contract. As we can see in the chart, the percentage of people who signed contracts did trend up, so the results of this test were good. We will continue to offer value propositions at the end of the tour.

In week 4 we tested not giving a tour at all. In this example I gave two choices: the amount of contracts went down or did not change. Depending on how the test went in real life, we could decide if this test was validated or not.

Cohort testing, where you compare one group of customers against another, is a tremendous tool to validate learning.

Getting better at getting Lean

Using MVPs and cohort testing will immediately bring your business a huge amount of value. I think enough to make it worthwhile spending the time and energy switching your mindset into the lean mode of thinking. Once you’re moving and testing, there is plenty more to learn. I recommend Eric Riess’ original book Lean Startup. A quick search on Google will turn up endless amounts of information about testing and other theory.

You can see my syllabus for Lean Startup here:

You can also see my advice for a startup (CoursAvenue) starting to go lean:

And if you have specific questions, leave them in the comments and I’ll respond. Good luck going lean!

Originally published at on April 18, 2013.