The Startup Way: How Entrepreneurial Management Transforms Culture and Drives Growth

by Eric Ries

The more difficult part is the probability-weighted distribution of future outcomes: the experiment.

When they experiment, they both reveal how large the impact could be and also increase the probability of it happening—and

It’s a measurement of what the startup has learned about far future profits.

company’s long-term growth and success, it creates a much closer alignment between the financial incentives of employees and managers and the organization’s long-term health. I

least distortionary set of incentives. It allows employees’ intrinsic creativity, commitment,

trailing indicators

leading indicators

Business plans tend to be made up of forecasts and predictions, always denominated in gross metrics (what we call in the Lean Startup movement vanity metrics

“no business plan survives first contact with customers,”

Steve Blank

There can be no question about how well the company is doing when everyone shares the same set of facts.

As I facilitated a Q&A between the two groups, one particular question was asked again and again by the executives: How do your investors hold you accountable? How often do you report progress to them? And how do they make sure you don’t go off the rails and do something stupid with their money? The founder/CEO of the startup was baffled by these questions. As it happened, I was an investor in the company. The executives were aghast that I’d allow him to spend my money without explanation or oversight. This was my chance to explain

we do risk mitigation in

metered f...

entitlement ...

validated learning,

Good ideas can come from anywhere, and people should be given resources and attention based on their talents, not their pedigree.

there is ample academic research showing that companies that believe themselves to be meritocracies are prone to more implicit bias than those that do not.16

What meritocracy actually means to Silicon Valley is that your credentials or qualifications don’t necessarily predict whether you’ll be a good founder or not.

there’s no one “right” leading indicator.

Meritocracy is not an either/or concept. Meritocracies exist on a spectrum. All of us can get better and become more meritocratic. And yet how many organizations truly live up to this ideal?

STARTUP IS MISSION—AND VISION—DRIVEN

Without a vision you cannot pivot. The accuracy of that statement is baked into the very definition of pivot: A pivot is a change in strategy without a change in vision

As Jeff Lawson, the CEO of cloud communications company Twilio, says, “You’re not going to get anywhere if you have a big vision but you’re not solving the customer’s problem. If you’re not solving a problem, you’re never going to be given the ability to implement that grand vision.” And the way to solve problems is to uncover them as you go and then pivot to meet them.

Genius is widely distributed, but as of yet, opportunity is not.

Identify the beliefs about what must be true in order for the startup to succeed. We call these leap-of-faith assumptions.

Create an experiment to test those assumptions as quickly and inexpensively as possible. We call this initial effort a minimum viable product.

Think like a scientist. Treat each experiment as an opportunity to learn what’s working and what’s not. We call this “unit of progr...

Take the learning from each experiment and start the loop over again. This cycle of iteration is called the bu...

On a regular schedule (cadence), make a decision about whether to make a change in strategy (pivot) ...

A simple way to do it is to make it a habit to write down expectations about interactions with a customer or colleague. “I believe the customer will be willing to participate in a follow-up call,” or “I believe this software feature will be important and appealing to the finance department at my company,”

What assumptions would have to be true in order for the project to succeed? Are they assumptions about customers? Partners? Competitors?

How much do we really know about customers’ habits, preferences, and need for solutions like ours? What evidence is there that customers really have the problem being solved for them and strongly desire (and are willing to pay for) a solution to it? What is really known about what customers want in that solution?

validates key assumptions: 1. Do people really have the problem you think they do? 2. How do they approach the problem today? 3. Is your concept a better alternative for them?

These surprises often upend our entire LOFA framework. That’s yet another reason why the LOFA analysis process should be kept as simple as possible.

We have the words Done is better than perfect painted on our walls to remind ourselves to always keep shipping.”

Teams that drive down the validation cycle time are much more likely to find product/market fit,6 because it increases (not guarantees, of course) the probability of success.

“The thing about Minimum Viable Products is that while you decide what’s Minimum, the customer decides if it’s Viable,” writes David Bland, a consultant and early Lean Startup evangelist.

Minimum Viable Products are optimized for learning, not for scaling. This is one of the hardest things to convey to people who’ve spent their lives building to build, not building to learn.”

“Go Broad to Go Narrow.”

He asked each person to consider the project she or he was working on at that moment and to spend five minutes writing down leap-of-faith assumptions.

But what constantly surprises me when doing this exercise with teams is how often startups are considering multiple MVPs that test exactly the same assumptions and yet cost dramatically different amounts of money. In these cases, we can almost always eliminate the more expensive MVPs from consideration, even if doing so feels uncomfortable.

Not figuring out how to achieve the specification with slightly less effort, but figuring out how to achieve the same learning value with a dramatically simpler specification.

validated learning

exchange of value

Validated learning is the scientific inference we can make from improvements in this exchange of value from one experiment to the next. (We’ll talk more about metrics in Chapters 6 and 9.) For metrics to support a valid inference, they must follow the three A’s: actionable, accessible, and auditable.

Paradoxically, when the quest for perfection is replaced with a willingness to experiment with and adapt the original idea, the ultimate result is a more perfect product. The key is to always drive down the total time through this loop. TWILIO’S BUILD-MEASURE-LEARN

PRFAQ,

What evidence do we have that our current strategy is taking us closer to our vision?

their funding—called a growth board11—not only saw that but applauded it.