The Digital Transformation Playbook: Rethink Your Business for the Digital Age

particularly clear to companies already steeped in the practice of running 
experiments. One survey of experiment-focused businesses reported that 
two-thirds of the new ideas tested by Microsoft failed to deliver any of their 
expected benefits. Only 10 percent of Google’s experiments were success-
ful enough to lead to business changes. And Netflix has estimated that 90 
percent of what it tries turns out to be wrong.
38
As technology journalist Alexis Madrigal has observed, “It turns out 
that our creativity is good but our judgement is lousy.”
39
There is a solution. Companies can compensate for the fallibility of 
management’s own judgment if they instill in their employees a culture of 
testing and learning about every aspect of their business. One company 
that has done so is Amazon. We can see this in the experience of Greg Lin-
den, a former Amazon developer. He was working on Amazon’s checkout 
process when he came up with the idea of offering shoppers a final set of 
product recommendations as they checked out, based on the items that 
were already in their shopping cart. When he presented the idea, senior 
management hated it. It was a cardinal rule of e-commerce to not distract 
or get in the way of the shopper once they have begun the checkout process. 
But Linden kept thinking about how checkout shelves in real-world super-
markets are ideal for getting customers to pick up just one more item on 
their way out. Although he had been forbidden to work further on the proj-
ect, he went ahead and built a quick test version of the feature. The senior 
vice president who had voted down his idea couldn’t have been happy, but 
the company let Linden run the test anyway. (At Amazon, it was hard for 

---

160
I N N O V A T E B Y R A P I D E X P E R I M E N T A T I O N
even a top executive to block a test experiment.) The data came back, and 
Linden’s innovation turned out to be extremely profitable. Resources were 
immediately applied to developing and launching a full version of it.
40
In how many companies would Linden’s story have ended this way?
Leading Without Deciding
The antithesis of Greg Linden within the world of retail might be Ron John-
son. In 2011, Johnson left Apple to take over as CEO of struggling retailer 
JCPenney. Johnson had a bold vision to reinvent the discount department 
store with a more modern, Apple Store–like environment. The retail experi-
ence was to be transformed—featuring smaller shops within the store, cool 
coffee bars to hang out in, and new outside brands like Martha Stewart. 
Eventually, all cash registers and checkout counters would be replaced with 
high-tech product-tracking and self-checkout systems. Johnson pledged to 
reinvent pricing as well, shifting from heavy use of coupons and sales pro-
motions to reliance on standardized pricing year-round. It was a truly bold 
hypothesis, but would JCPenney’s customers respond positively to a radi-
cally different type of store? Unfortunately, after years of success leading 
retail teams at Apple, Johnson felt no need to test his hypothesis. Instead, 
he simply rolled it out, with no pilots and no limited test markets. The 
result was a catastrophe. The company, which had already been suffering 
for years, fell into much steeper decline. A little after a year under Johnson’s 
leadership, its quarterly results showed a 32 percent drop in same-store 
sales—what some observers suspected was the worst decline ever reported 
by a major retailer in history.
41
Seventeen months into his tenure, Johnson 
was ousted as CEO.
One can only imagine what might have transpired if Johnson had 
instructed his team at JCPenney to test the assumptions behind his new 
strategy in a series of early and focused experiments. Rapid experimenta-
tion requires more than curious and empowered employees like Linden 
in the trenches; it requires a different kind of leadership from the top, too. 
Nathan Furr and Jeff Dyer talk about this as a shift in role from “Chief 
Decision Maker” to “Chief Experimenter.”
42
In the experiment-driven 
organization, leadership becomes less about making the big decisions on 
behalf of the organization. The role of a leader, whether CEO or head of 
a small team, shifts from providing the right answers to posing the right 
questions.

---

I N N O V A T E B Y R A P I D E X P E R I M E N T A T I O N

161