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

Why Digital Is Impacting Both
Digital technologies are making rapid experimentation both more possible 
and more necessary than ever before. They are offering new tools for exper-
imentation and increasing the speed at which companies must innovate to 
keep up with a rapidly changing environment.
Convergent experimentation is becoming increasingly powerful 
and affordable due to new technologies. As companies in every industry 
develop digital products and services for customers (and processes for 
employees and partners), these digital innovations are inherently much 
easier to test in real time and at low cost. (Think of how much easier it 
is for a bank to test the design of its mobile app than to test the design 
of its retail branches.) At the same time, new software tools are becom-
ing available that allow even small firms with limited budgets to easily 
conduct A/B tests, run multivariable analyses on the results, and deter-
mine the optimal sample size for an experiment. Optimizely, a start-up 
cofounded by one of the early experimenters for the original Obama cam-
paign, allows small businesses to start running A/B tests on their web-
sites and mobile apps for free. The increasing focus on data analytics in 
companies of all sizes is making convergent experimentation widespread 
across industries.
As digital computing becomes more ubiquitous with mobile comput-
ing and the Internet of Things, the possibilities for convergent experiments 
will only increase. Imagine a grocery store wanting to test four possible 
promotions for its store-brand barbecue sauce. In the analog age, it would 

---

134
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
have needed four sets of stores, each running a different promotion. But, 
today, if it can use mobile or wearable devices to push the promotion digi-
tally to consumers, even a single store could test four versions with random 
selections of customers in that store.
Divergent experimentation is gaining new tools from digital technol-
ogy as well, particularly in the form of new ways to prototype ideas cheaply 
and rapidly to show customers. For new physical product offerings, both 
3D printing and computer simulations decrease the time and cost involved 
in creating prototypes. For digital products and services, newer program-
ming languages and repurposable code make it easier to develop “good 
enough” prototypes to test with customers. Even in industries like pharma-
ceuticals, as robotic systems take over the purely manual tasks previously 
done by junior lab technicians, the ability to rapidly and cheaply test new 
molecular and genetic combinations is increasing dramatically.
In the digital age, even the biggest companies are striving to innovate 
faster and become more “agile” and “lean” like start-ups. Fortunately, thanks 
to digital tools, all companies are able to run more experiments—both 
convergent and divergent—cheaply and quickly and accelerate the pace 
of innovation. As technological change continues to impact every indus-
try, experimentation will become more important than ever as a means of 
reducing uncertainty and accelerating innovation.
Seven Principles of Experimentation
Applying experimentation to a business is not easy. To create the most 
value for your innovation efforts, a few principles are critical. These have 
been identified by observing innovative companies in a range of industries 
and by surveying the leading research on innovation from the past decade. 
These seven principles apply for any business experiment, whether conver-
gent or divergent:
r
 Learn Early
r
 Be Fast and Iterate
r
 Fall in Love with the Problem, Not the Solution
r
 Get Credible Feedback
r
 Measure What Matters Now
r
 Test Your Assumptions
r
 Fail Smart

---

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

135
Let’s take a look at each.
Learn Early
The first principle is to start experimenting from the very beginning of 
your innovation efforts so that you can learn as early as possible in the 
process. The same lesson that would trigger heavy financial loss at the end 
of a product development process (“our customer didn’t need this solution 
or wouldn’t pay for it”) can come fairly cheaply if it is learned in the early 
stages of your project. You can think of this effect as “the value of early 
learning” or, conversely, “the cost of late learning.” (See figure 5.1.)
Hanson described this phenomenon in terms of the shift at Intuit from a 
traditional innovation process—in which customers are exposed to a prod-
uct only after a long design and development stage—to a process of rapid 
experimentation—in which customers are brought in much earlier to pro-
vide the feedback that helps the company decide which ideas are even worth 
pursuing. With much earlier learning, the failure rate for the company’s 
product ideas did not decrease, but the cost of failure dropped dramatically. 
“We can run 50 different ideas through our rapid experimentation process 
in the time and resources it takes to run 3 ideas through our old process.”
11