Book review: The Lean Startup — Eric Ries
Published in 2011, the Lean Startup has been considered the startup bible for almost a decade. I wondered how the book stood its test of time, and re-read it a few weeks ago. The only thing I remembered from reading it the first time was its famous “build-measure-learn” loop, therefore I was surprised the book is more about entrepreneurship in general, management & accounting, than about MVPs (minimal viable products) and customer interviews. It turns out the famous lean startup method is about more than building MVPs as fast as you can. My 5 take-aways:
“A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty”.
As a lot of writers before him, Mr Ries starts with defining the subjects of his book: entrepreneurs and startups. He takes as broad a view as possible, as a lot of his book is on internal entrepreneurs (intrapreneurs) as well, and this is his way to incorporate everyone in the definition. I agree with his view of a startup, but disagree on what he defines as conditions of extreme uncertainty. Eric Ries applies this uncertainty on the product and market specifications: what should the product be, the price, the market, the business model, etc and in that way sets the startup apart from the bakery around the corner, but at the same time includes the corporate innovation teams.
In my view the corporate innovation team is doing something different than the startups. For startups the uncertainty extends to other areas as well: where should our employees come from, how much should we pay them, what should we pay ourselves? Do we still have a company next week? Will I be able to pay my rent next month? And the hardest one of all: to pivot or to persevere? This is the uncertainty which defines a startup just as much (or maybe even more) than not being sure about the revenue model of the business.
“Learning is the essential unit of progress for a startup”
Mr Ries answers one of the biggest questions of any incubator or accelerator program. If traditional metrics like revenue or gross margin do not apply to define progress of a startup, then what does? Amount of FTE? Funding? # of customers? If the goal of a startup is to create certainty, then the progress metric should be exploring the uncertainty, learning about the uncertainty, and like any proper explorer before you, slowly changing the uncertainty to certain. But how, Mr Ries asks himself, do you measure how much you learn? And more importantly, what is most important to learn? And does it matter how fast we learn? He answers most of these questions himself: we measure how much we learn by seeing learning as a scientific experiment, which means: draft a hypothesis, design an experiment to test, build the experiment, perform it, and draw your conclusion. It turns out there are only two hypotheses worth testing: the value hypothesis and the growth hypothesis. And yes, the faster you learn, the better.
“If we do not know who the customer if, we do not know what quality is”
This one ties into the first hypothesis: is the team able to deliver value to the customer and accidentally coincides with a very human issue: the need to deliver quality. The faster the startup goes through the Build Measure Learn loop, the better. Speed can cause sloppiness and introduce errors. On the other hand: as a professional it does not feel “OK” to ship out a product of which you know you can do better. Or even worse: you know it will fail in multiple ways. Will this faulty product not harm your career? Will customers disproof of your product because it lacks features XYZ or does not look good enough?
Startups have been plagued by these discussions since the beginning of times and the solution for these discussions is rather simple: it is not up to the team to decide what quality is, what kind of features the product should need and what kind of colourways should be available. Only your customers are able to define value, quality and features and they are only able to do so when they use your product or service.
Only if the quality gets in the way of learning (your product won’t start, so i can’t use it, therefore I can’t give feedback on the product) it should become an issue. This means every professional should step over his or her high standards and accept that they will ship “poor” products, include features that they do not like, or exclude features that they absolutely love! Keep in mind that in the early days the goal is not to produce the best product out there, the goal of the startup is to learn as fast as possible.
(Don’t assume poor products and poor internal processes are the same by the way. You need very strict internal processes to deliberately ship “poor” products).
“How confident are you that you are making the right decisions?”
If the question above ties to value (and therefore how you approach your customer) this question ties into your internal processes and the second hypothesis coined by Ries: the growth hypothesis.
Imagine this: Your startup is doing great. New customers keep coming in, revenue is going up as well and you just received a new round of funding. But at the same time you’re spending a hell of a lot of time discussing improvements, updates & add-ons. Product updates, updates on your website, new sales processes and improved customer support are just some of the things on the table And if you’re perfectly honest, you don’t know where your time is best spent, so you do what you did before: you spend your time improving every bit of the company, it seems to work, right?
This scenario should sound familiar for almost every startup out there. The answer is, according to Ries, to start using actionable metrics & the earlier mentioned experiments. Here it gets tricky as a growing startup normally has a growing number of users, and it is very tempting to use this number to give yourself positive feedback. However: you need to have a running growth engine, which basically means you should improve every step of the way, and if your efforts are not improving the company, your efforts should be directed to some other part of the company.
Actionable metrics and cohort analysis (is this month group of users behaving different from last month group of users, instead of “he! We gained another 128932 users this month”) will help you to tweek your growth engine. This seems easy, but I’ve basically seen every startup struggle to bring actionable metrics in practice. Cohort analysis is relatively simple, but will start showing a world which you (and your investor) might not want to see!
“On the one hand […] we should get a product into customers’ hands asap, on the other hand […] shortcuts taken in product quality, design or infrastructure today may wind up slowing down the company tomorrow.
Engineers turned entrepreneurs often struggle with this balance of scaling and delivering quality at the same time. The question here is actually the same as in the issues mentioned above: how do you decide if the investment is worth the benefit of speed? This is a hard question as both the investment and the benefit are unknown quantities and this question, again, gives a lot of discussion around the management table.
The answer according to Lean Startup is to start asking the 5 Whys (link to 5 whys). The 5 Whys forces you to consistently make a proportional investment at each of the levels of the Whys. This sounds a bit cryptic, but it basically means you end up fixing things only up to the level they are broken. In that sense the approach acts as a natural speed regulator: nothing broken? No issues? Full throttle. A lot of issues? Slow down and fix them. I’m not a big fan of this kind of “harsh” methodologies, but this framework does give you a simple technique to keep growing while keeping your quality high.
What I’ve learned from Lean Startup
Startups are defined by their uncertainty. It is this uncertainty which makes them interesting and valuable. Unfortunately, humans are very badly equipped when it comes to uncertainty. The uncertainty often leads to endless discussions in startups. This endless fruitless discussion is a huge waste of time, and as speed is your ally in a startup it’s even worse!
Lean Startup hands you a way of working based on assumptions and experiments. Coupled with rigorous accounting on your customers behaviour this leads to a clear development path as a company, which is not only valuable because it saves time, but mostly because it gives peace of mind (and less discussions ;)). Instead of luck, or gut feeling, your startup is based on experiments, which will let you know, in time, if you’re growing in the right or the wrong direction.
