A bluffer’s guide to The Lean Startup method


Lean Startup

Eric Ries published hit book The Lean Startup in October 2011. Two years later, everyone in the startup world should be able to recite his incredibly useful methodology even if woken up with a bucket of cold water at 3am.

No? Here’s a short guide to the main concepts from Thomas Janson, who’ll be hosting a VentureVillage and Gründerszene Workshop on exactly this topic on 22 November 2013. 

The ultimate goal is to find product-market fit

The Lean Startup methodology is a three-layered model made up of the vision, the strategy and the actual product. It is also two-phased. The first phase is focused on validated learning and how to create the best value for the consumers of the target market as well as how to best grow.

Only once this is achieved – a state Ries calls product-market fit, including a working engine of growth – does the actual growth phase start. The goal is to find this fit and make a startup’s cash flow positive and sustainable before it runs out of money.

Keep the initial vision broad

Until a precise business model is found, it’s important to keep the vision of gaps and shortcomings in the market broad. This way, adjustments can be made to the model without reassessing the entire market.

An easy example? Bringing a new service to a two-sided market, making a wrong value assumption and charging the wrong group. In this case, it’s a good idea to be agnostic as to who your actual paying customers will be. You might have to switch from a paid to free model and create a new type of paying customer: advertisers.

The Lean Startup gives 10 examples of these types of changes – well worth studying. Flickr, for example, started as an online role-playing game and later switched to a photo community when it realised that’s where its users were most active.

Strategy = business model

The strategy is the actual business model. Like the initial vision, the initial strategy is based on the entrepreneur’s knowledge and skills. It can be defined by:

  • A value hypothesis: What product or service will satisfy that demand?
  • A growth hypothesis: How can growth be made sustainable so that it happens without requiring new funding?

Any hypothesis risks not holding true when put to the test. How to test a hypothesis? This can be illustrated with case studies. Apple didn’t have to worry that users would be reluctant to wear the iPod in public because it was just a Walkman without changeable media. The innovation, the bit that really needed to be tested: the paid download and synching part. Those uncertainties, Ries says, are leap-of-faith assumptions. Learning should be focused on validating them.


Start with a Minimum Viable Product (MVP)

The initial product is not yet steered by the Lean Startup methodology – at this point, it’s a rather non-scientific best bet at what could satisfy a demand. Here’s Ries’ take on it:

  • At this point, the goal is learning quickly not growing quickly. The first product is therefore not necessarily required to be something that can easily scale.
  • Everything that doesn’t aid the validated learning process is waste and should be cut.
  • Early adopters don’t necessarily care about polished products (or might even despise them) but rather about innovation and getting involved.

Combined, this allows for the building of a so-called minimum viable product that is specifically set up to be an experiment to validate leap-of-faith assumptions

Tune the engine (testing, testing)

The MVP needs to be tuned through validated learning to get a product/market-fit. Figure out users’ approval not by asking them for opinions but through real usage (time spent with a service can be a value statement) or purchase decisions. One example: A small online shop that asks for real payment data entry – only to tell users trying to buy that the product is not yet available. The point of this smoke test is to test whether users will actually try to buy a product, which is much stronger than just asking them for an opinion.

The results of the first MVP iteration will never meet expectations. Constant improvement is required. After the first-run, make a small change and test against the initial result using cohort or funnel metrics – in contrast to vanity metrics (the “hockey stick” growth curve).

The idea is to test for current actions not just accumulation or higher ad spending. These recurring measured improvements are called the build-measure-learn feedback loop.

If the team runs out of techniques and ideas to further improve the product and the results aren’t yet close to market potential, a pivot to a different business model might be needed.

The three engines of growth

Ries finishes with how best to think about growth. His first mantra: “new customers come from the actions of past customers”. This leaves room for three different engines of growth:

  • Sticky: Churn rate < organic (unpaid for) growth
  • Viral: One new user (virally) leads to more then one additional new user
  • Paid: Acquisition costs per user < lifetime value per user (creating the funds for funded advertising)

His second mantra is that in most cases, growth will come from one of these three engines and relying on two or all of them may indicate a weak model or a missing focus.

A note of caution

This is only a very short overview for the The Lean Startup’s radically analytical approach. To end with a quote from Scott Case, CEO of the Startup America Partnership: “Every founding team should stop for 48 hours and read The Lean Startup. Seriously, stop and read this book now.” Or just come to the workshop.

Image credits: 
Lean Startup books: Flickr user Betsy Weber
MVP FTW: Flickr user Fake Grimlock

[contentad keyword=]