Author: Eric Ries
Started reading: 18.June.2016
Finished reading: 21.July.2016
There are two broad ways of thinking about the founding of a company.
Some believe that a startup is a moment in time.
Watch movies like Jobs or The Social Network, and you’re in no doubt that a company is an original dish, conceived on the palate of a chef extraordinaire.
But IMVU co-founder Eric Ries believes it’s more like a recipe that can be learned by anyone.
The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses reads too much like a manual but offers a heap of practical recommendations for startup entrepreneurs.
The ‘Lean Startup approach’ is organised around the principle that a startup company’s activities should exclusively be geared toward generating ‘validated learning’, empirical data about the value of a product. Anything else is ‘waste’.
Why so serious? The reason is that startups operate in a context of uncertainty. As Ries says, “in a startup, who the customer is and what the customer might find valuable are unknown”. If the product doesn’t ‘fit’ the market, and no one is interested, time and resources have been wasted for nothing.
This was Ries’ experience at IMVU, a tech startup that had a shaky beginning before finding its feet. Its first product, 3D personal avatars which customers could assign to their existing instant messaging accounts, was a flop. Its current iteration as a 3D chat community has over 3 million users.
In ascertaining whether customers will actually use a product, it’s important to note that asking customers what they want doesn’t cut it as research.
Customers don’t necessarily understand their own preferences. If Steve Jobs had asked Americans in the 90s what they wanted, no one would have answered, “the iPhone” or “an iPad”. And someone saying they ‘like’ a feature is very different to them putting money on the line and buying it.
So how does a startup learn quickly and without waste?
The only answer, Ries argues, is to build a minimum viable product (MVP), release it, gauge how people respond, then decide whether to pivot or persevere.
Two further tools are important here:
> split-testing (to determine the most useful qualities of the product). This sees you conduct randomised control trials of your own product or webpage, using services like Google Analytics. This is also known as A/B testing.
> small batches (to minimise the quantity of the product, and therefore eliminate the risk of waste). Taken from ‘lean manufacturing’ companies like Toyota, the notion of working in small rather than large batches seems less efficient but enables a company to discover and rectify mistakes (or crappy products) without wasting larger amounts of time and money.
Definitely read this book before starting a company (in case you’re about to blow your credit card for nothing).
> Collect granular data, don’t fall for vanity metrics. Metrics like ‘customer growth’ might seem promising but don’t tell you anything about the relative identity or value of the users. Are you retaining them? Are they referring their friends? Are you enjoying a temporary uptick in growth thanks only to an effective marketing campaign? A website is getting 20, 000 hits per month. Does this represent genuine interest? Or are there only a few, highly active visitors? In contrast to vanity metrics, cohort analysis (breaking data down into sub-groups of people) provides more detailed and usable insights.
> Start small, not perfect. Don’t fall into the trap of thinking a product needs to be in its perfect, fully-featured, final form before you release it. Launch a MVP and test it out before gambling the house.