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

Can limit
roll-out
Cannot limit
roll-out
Polished
launch
MVP
launch
Figure 5.4
Four Paths for Scaling Up.

---

156
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
MVP Rollout
This is the easiest path for introducing an innovation because you can start 
your rollout with a limited test market and then iterate rapidly as you gain 
additional feedback from customers. In these cases, you may bleed right 
from your minimum viable prototype into actual product development. 
That is, your first public release will be a minimum viable product offered 
to a limited set of customers. The relative ease of this path is one upside to 
being a little-known start-up: you can iterate and learn with real customers 
without much public scrutiny.
This was what Rent The Runway did after receiving its first round of 
capital from Bain. The first website launched with only 5,000 members, 
by invitation only. This allowed the company to start with a relatively 
inexpensive inventory of 8,000 dresses from thirty designers. Once they 
saw the business model was succeeding and press coverage led to a surge 
in requests to join, the founders secured a second round of financing so 
they could scale up quickly to meet demand. An example of a locally lim-
ited MVP rollout is the launch of Zipcar. This was one of the first services 
to allow members to rent a car by the hour, picking the cars up at street 
locations identified online rather than having to visit a car rental office. 
Founder Robin Chase launched Zipcar as an MVP only six months after 
beginning work on the business and having raised just $75,000. She was 
able to do this partly because she began only in Boston, waiting more 
than a year to extend to a second location. This allowed her to test out 
the business model and iterate her service with feedback from paying 
customers.
MVP Launch
The second path for scaling up is harder. In this quadrant, your business 
is forced to iterate very quickly after launching your innovation because 
you are not able to able to effectively limit the scope of the launch. (As 
a result, your first release could make a lasting impression on a larger 
audience.)
One reason this path may be necessary, even for a digital service, is 
that the business has to rely on network effects. For example, eBay was 
predicated on a platform business model that required both buyers and 

---

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

157
sellers. Growing each side of that equation as quickly as possible was 
essential (no one wants to sell on an auction site with few customers 
or to browse on an auction site with few products). Omidyar could not 
afford to restrict the website to a small pool of customers while he iter-
ated and perfected it.
A business also may not be able to limit the release of an innovation 
due to the high visibility of its brand or the expectation that the initiative 
may draw wide attention. American Express launched Small Business Sat-
urday with the idea of putting a spotlight on America’s small, local busi-
nesses for one day. The campaign launched in just six weeks with its scope 
still undetermined. An outpouring of energy and involvement came in 
from social media, consumers, business owners, and even an act of Con-
gress. The company had to move quickly, but it was able to rapidly evolve 
the program and its goals as Small Business Saturday quickly became an 
annual phenomenon during the holiday shopping season.
Polished Rollout
The third path for scaling up is also harder than the first—but for different 
reasons. In this quadrant, you are able to launch your innovation in limited 
locations or for limited customers, but you cannot quickly iterate it once 
it is public. It therefore needs to be much more polished at the point of 
release.
Still, you are able to take advantage of rolling your innovation out in 
stages by validating your initial findings and testing how it is received by 
different customers or in different markets. Retail design typically follows 
this path. Starbucks has tested diverse ideas, such as offering local wines 
and craft beers, in a set of store locations in Seattle. The company first tested 
wireless charging mats for phones at stores in Boston before rolling them 
out nationwide. It even tested a coffee delivery service (via mobile app) by 
making it available exclusively to customers working in New York’s Empire 
State Building.
When Settlement Music School, an education nonprofit in Philadel-
phia, developed an innovative plan for a new music program aimed at 
adults, it chose to roll it out in one location at a time. After the first two 
locations succeeded but the third foundered, the school realized the pro-
gram would need to be adapted based on the musical interests and cultural 
networks of each surrounding neighborhood.
36

---

158
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
Polished Launch
The fourth path for scaling up a new innovation is the hardest of all. In this 
quadrant, you must offer your new innovation to all customers at once, and 
you are unable to iterate it quickly. This creates maximum pressure for your 
company to polish and carefully test an innovation before its public release.
This is the path for innovations like new automobiles, pharmaceuticals, 
and hardware products. In cases where a physical product can be updated 
in a year or less (e.g., some consumer electronics), you may want to aim 
for a streamlined first product, withholding some of your eventual fea-
tures until the first edition is on the market. This is the pattern of Apple’s 
most successful products, which typically have made large leaps in features 
between their first and second years (in that sense, some would say the 
first-generation iPads and iPhones were both “MVPs”).
By contrast, we can look at Google Glass. The wearable eye-frame com-
puting device was released publicly while it was still buggy and before Google 
was even clear on the value proposition for the user. The company failed to 
iterate Glass meaningfully within a year because it was still just trying to get 
the device to work consistently. It was probably used to operating in the MVP 
rollout quadrant (where it had launched Gmail and countless other software 
products), and it underestimated the discipline necessary when releasing a 
hardware product, especially one that would be attracting massive media 
attention. Although Google released Glass to only a few thousand custom-
ers, the prominence of its brand and the controversial nature of the product 
(with its ability to record video incognito) ensured that the release was sub-
ject to prolonged and intense scrutiny. A national conversation ensued about 
what Glass meant for the future of computing and privacy, and the company, 
which grew up with the most casual of beta-style launches, learned that not 
every new innovation can be released the same way.
Knowing which of these four quadrants your innovation fits in—
polished or MVP, rollout or launch—will clarify your path to bringing it 
forth and scaling it up successfully. Any new innovation should continue 
to iterate and improve after launch. Knowing how to best do so is essential.
Organizational Challenges of Innovation
Putting rapid experimentation at the heart of the innovation process is not 
easy for many large or traditional organizations. As they have grown, most 
businesses have relied on decision making by committee or by seniority 

---

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

159
and chain of command. In Silicon Valley, it is commonly said that HiPPOs 
make the decisions at more-traditional firms. (No, not the river-dwelling 
mammal you see in the zoo. This is decision based on the Highest Paid Per-
son’s Opinion.) Rethinking innovation requires significant organizational 
changes, beginning with how decisions get made.
Building a Test-and-Learn Culture
Historian Yuval Noah Harari describes the birth of the Scientific Revo-
lution as “the discovery of ignorance.” In his view, the birth of modern 
human societies began with this credo: “We don’t know everything … the 
things that we think we know could be proven wrong … no concept, idea 
or theory is sacred and beyond challenge.”
37
For a business to embrace experimentation requires a similar rec-
ognition: we do not know what we think we do. This sobering truth is