The Three Stages of Disruptive Innovation: Idea Generation, Incubation,

and Scaling

1. ideation, to generate potential new business ideas;

2. incubation, to validate these ideas in the market;
3. and scaling, to reallocate the assets and capabilities
needed to grow the new business.
A McKinsey study found that 70% of the sênior executives surveyed
listed innovation as a major concern. And no wonder they are worried. In
the face of these seismic shifts, companies around the world have
embraced a variety of efforts to drive innovation, including everything
from setting up outposts in Silicon Valley, to launching corporate venture
capital funds, to embracing design thinking and the lean startup
methodology, to open source innovation.

Given the energy, time, and money spent on these innovation

efforts, why have the results not been more impressive? Why have more
companies not cracked the code of disruptive innovation? One important
reason is that firms have not appreciated the need to manage
discontinuous and incremental innovation differently. Incremental
innovation enables companies to do the important work of developing
new products, extending the life of existing ones, refining existing
processes to become more efficient, and finding new customer segments
to drive revenue growth. These efforts are incremental innovations that
exploit existing assets and capabilities. In contrast, discontinuous
innovation helps firms to develop new capabilities and assets, often selling
to new customer sets.

Unless there is a clear strategy justifying the entrance into the new
business, and unless senior management is prepared to protect these
embryonic efforts, the tendency is for the mature business to either starve
the new business or to impose on it the performance standards of the
mature business, an easy way to kill the new venture.

Driving disruptive innovation in large companies requires firms to be

ambidextrous— to compete in mature markets where efficiency, control,
and incremental improvement are essential (exploitation), and to
simultaneously compete The Three Stages of Disruptive Innovation: Idea
Generation, Incubation, and Scaling in new technologies and markets
where flexibility, autonomy, and experimentation are needed
(exploration). This requires mastery of three distinct innovation
disciplines: idea generation or ideation or the discovery and development
of ideas for potential new businesses; incubation, where the new ideas
are validated in the market; and scaling, where existing assets and
capabilities are reallocated to help the new venture grow. Managers have
too often concentrated on the first two of these disciplines and have
neglected the third.

Below, we review and discuss the tools and methodologies required

by each of the three disciplines required by ambidexterity and illustrate
how idea generation or ideation (e.g., design thinking, corporate venture
capital, open innovation, and hackathons) is different from incubation
(e.g., business design experiments, lean startup, business canvas), and
how these are different still from growth or scaling. While each of these is
necessary, unless all three disciplines are mastered, none will be sufficient
by itself. Some firms have developed processes that successfully leverage
all three disciplines, while other firms that have concentrated on only one
or two of them have struggled or failed. Interestingly, an examination of
the successes and failures suggests that the crucial underlying issue is not
technology or organization design but leadership.

Ideation: Generating Ideas for New Businesses:

The following four such approaches have been widely adopted:

1) Open Innovation: These include opening the company’s platform and

sharing intellectual property so others can develop products with it
(e.g., Apple, Google, Amazon, and Intuit), inviting customers to suggest
ideas for new products and services (e.g., Lego, P&G), offering contests
for participants who can provide solutions for company problems (e.g.,
NASA, InnoCentive, and Kaggle), or fostering networks or communities
of interest (e.g., IBM, Wikipedia, and Facebook).
2) Corporate Venture Capital: Another common approach is for
companies to develop relationships with startup ecosystems through
corporate venture capital units with the objective of gaining insight
into the innovation efforts of startups. Unlike traditional venture
capital where the objective is to fund new businesses and generate a
financial return, the objective of corporate venture capital is typically
to identify and exploit synergies between the startup and the larger
firm and provide opportunities for growth. Firms may invest in startups
to provide a window on new technologies, to explore new business
models, or to enter new markets. The startup company may get
financial capital, access to channels and customers, and the expertise
of the larger firm (e.g., technology, manufacturing, and distribution).
Often, these investments are made with an eye to a future acquisition.
Like open innovation, these investments provide a way for established
firms to access ideas outside the boundaries of the firm.
3) Design Thinking: Design thinking is a methodology for stimulating
creativity—a way of generating insights into the real problems faced by
customers and rapidly generating prototypes or potential solutions.
Originally developed by David Kelley and the design firm IDEO, design
thinking seeks to release the human capacity for creativity that can be
stifled in mature organizations. To rediscover this, he developed an
iterative process called design thinking that first encourages generating
new ideas and insights (creativity) through empathic listening and then
narrowing the focus through rapid prototyping and testing
(implementation). The process relies on the following principles (or
practices):
a. Empathize—Begin by deeply understanding your customers’
problems. This requires understanding your customers in their
environment and gaining empathy.
b. Define—Do not jump to solutions before you are clear about the
real problem the customer has. This means being open to
changing the initial definition of the problem based on your
insight into the root cause of the problem.
 c. Ideate—Use brainstorming (wild ideas, no criticism or
evaluation, build on the ideas of others, yes-and) to generate
alternative ways to address the customer’s pain point.
d. Prototype—Develop low-resolution prototypes of your solution.
Focus on prototypes that will test the key insights you have
about the customers. Do not let the perfect be the enemy of the
good.
e. Test—Share the prototype with the user and listen carefully for
their reactions. Use these to develop deeper insights into their
needs. Use these to iterate and redefine the problem, which may
lead to further insights.

The process of design thinking is a methodology for stimulating

creativity. It enhances ideation and provides some initial data about
customer acceptance. But it does not provide significant evidence about
the business value of the proposed solution (incubation) or whether the
underlying idea justifies taking assets and capabilities away from the
current business (scaling).

4) Employee Involvement: The role of employees in innovation has

evolved substantially beyond suggestion boxes to encompass online
suggestion systems, internal contests, and hackathons.

What is common across these various approaches is how they help

to surfasse new ideas, either from outside the company (open innovation,
corporate venture capital) or from within (design thinking, employee
involvement). Done well, they can enhance creativity and, critically, in the
case of design thinking, focus idea generation on high-value customer
problems.

These idea-generation techniques are agnostic about the nature of

the ideas that are generated—disruptive or incremental—but generally
promote incremental innovations. Unfortunately, there is no necessary
relationship between excellence in these practices and being better
positioned to take advantage of market disruptions. In fact, there is
evidence that the better firms are at promoting incremental
improvements, the worse they are at discontinuous innovation. They have
mostly been applied to develop products or processes that fit within
existing business models. Only rarely are they applied to solve the
problem of how to build a new scalable business. Firms then need to do
more than open innovation or design thinking if they are to generate ideas
of the sort required for disruptive innovation. There are two practices that
can help ideation produce ideas suitable for validation and scaling:

1. Scale of Ambition—Setting a scale of ambition that is equal to the

opportunity or threat of disruption helps to move ideation away from
the incremental or tactical. This means defining an aspiration for a new
business or business model, not just a new product or service—for
example, a technology firm declaring that they want to create new
revenue streams from services, not just sell existing components. For
example, at Amazon, ideas for new businesses are considered if they
meet three criteria: they offer a differentiated customer experience,
they can grow into a large business, and they can provide great returns
on invested capital.
2. Hunting Zones—Aside from setting ambitious goals, it is important to
put boundaries around ideation by defining the markets, business
models, types of problems, or customers to focus on. This ensures
effort is focused on áreas most likely to deliver on the ambition.
Otherwise, the democratic style of ideation techniques, like
hackathons, can result in a plethora of ideas that help exploit the
existing business but may not address disruptive threats. These
guardrails help ensure that ideas are considered and should include an
assessment of the attractiveness of the opportunity (market size, ease
of penetration, substitute threat, etc.).

If ideation techniques are focused at the right scale of ambition and

within well-defined hunting zones, and if the subsequent validation of
these ideas is based on research, then the probability of them meeting the
challenge of disruptive innovation increases. Without them, there is a
strong prospect of developing an “experiment zoo,” where people and
resources are fragmented across multiple potential areas of opportunity.
For example, a large Japanese electronics firm embraced ideation and
generated more than 400 new ideas, but, when pressed, acknowledged
 that only two of those had actually generated revenue, and these were
incremental additions to existing businesses. Ideation is a necessary but
not sufficient first step in ambidexterity.

Even with a well-run ideation approach, many, if not most, ideas

generated are not very good. That is the nature of ideation: to generate a
diversity of options. So, how does a leader separate the good ones that
are worth investing in from the bad? The answer is through incubation—a
process to determine whether the idea meets the market test.

Incubation: Validating the New Idea

Given the above activities, companies usually do not lack new ideas.
But how does a leader separate the good ones that are worth investing in
from the bad? The answer is through incubation or validation—a process
to determine whether the idea meets a market test. This is the second
discipline required for ambidexterity. Three useful methodologies address
this challenge: the lean startup, the business model canvas, and the
Stanford Launch Pad. Each of these offers a way for leaders to test their
ideas in the marketplace.

1. Lean Startup: The lean startup begins with an entrepreneur’s

hypothesis about a new product or service. The idea is to work backward
from the business results you are trying to achieve rather than forward
from some solution or technology you want to sell. The intent is to
eliminate wasteful or unnecessary practices, focus on quickly designing
and running an experiment to test the original hypothesis, and iterating
based on the results of the experiment. The approach emphasizes a build-
measure-learn logic with the development of a minimally viable product
(MVP), putting the product in front of potential customers, rapid iteration
and pivoting based on this learning, and the use of metrics that can lead to
informed decisions rather than vanity metrics that make the manager look
good. Although the methodology was originally designed to help
entrepreneurs, it has been enthusiastically embraced by large firms from
around the world.
2. Business Model Canvas: This approach provides a set of nine building
blocks that can be used in establishing a new venture. It helps users to
 systematically think through their business model and identify those
elements needed to test the original hypothesis. The canvas can be
adjusted as a company grows. Variants of the lean startup and business
canvas methodologies are often used by startup accelerators like Y
Combinator and Plug n Play. These approaches typically provide some
minimal funding and emphasize intensive interaction with customers,
rapid prototyping, and fast iteration.
3. The Launch Pad: This is a Stanford Design School class for entrepreneurs.
The class begins with an idea for a product or service and adresses three
questions: Who is the hyper-specific target user? What is their specific
pain point? and What single function have you performed to reduce this
pain? Over the course of ten weeks, participants are required to talk to a
minimum of 100 potential customers. Combining elements of design
thinking, the lean startup,
and the business canvas, the emphasis is on listening to the customer,
rapid prototyping, and fast iteration. Taken together, these approaches
have had a measurable impact on helping both entrepreneurs and
established firms incubate new ideas. While most often these are focused
on incremental innovation developing new products and services, they
can also be used to generate new businesses. However, these approaches
are largely silent with respect to scaling; that is, if the entrepreneurial idea
begins to grow, neither the lean startup nor the business canvas offers
guidance for how to design the organization to ensure that the growth
trajectory is sustained. Note that while the lean startup methodology
works, the larger problem is that big companies do not know how to scale
the new venture.

Hypothesis Testing—Central to incubation is the idea of an iterative loop

between an assumption about the market opportunity, actual experience
with customers that confirms or refutes that assumption, and adaptations
to the model based on learning. This approach places a high value on
learning through many small failures, not the route to success for most
corporate executives. The idea is not to build an entire solution to test but
to formulate a series of small tests of limited hypotheses. Jeff Bezos at
Amazon captures this tension, observing that “Most large organizations
embrace the idea of innovation but are not willing to suffer the string of
failed experiments necessary to get there.” For this reason, there is both
a lack of familiarity and comfort for many large firm executives with the
basic skills of formulating and testing a hypothesis with data. Yet, without
this discipline, the risk is that a business moves to scaling based on
unproven assumptions or a flawed experiment.

Feedforward Measurement—Another key area in which incubation

challenges the typical practices of a core operating business is the
measurement system. Most organizations review data on past
performance, compare it with expectations, and act to correct errors. This
is a feedback loop—What was our goal? How did we do? What explains
the variance, how can we close the gap? What is required for an
experiment is a feedforward system that tracks performance toward a
strategic goal. What data do we have to tell us how an experiment is
performing relative to its hypotheses? Similarly, when incubating new
ideas at Amazon, the emphasis is not on measuring outputs (e.g., revenue
or new customers), but focusing on assessing the inputs that drive the
outputs (e.g., speed of delivery or rate of adoption).

Executive Oversight—Senior managers need to be formally involved in the

decision making on experiments in the incubation process. The biggest
threat to moving from incubation to scaling a new venture is a profitable
business unit that opposes diverting current investments with certain
short-term rewards to the less predictable opportunity of creating a new
disruptive venture. These managers are not acting with malice. It is often a
rational choice to argue for certain gains over future possibilities. So,
when the moment comes to commit, there needs to be clarity and a
shared understanding about the ambition for the new venture. That
means senior executives must commit time and attention to reviewing the
experiment as it unfolds. Getting that level of engagement can be difficult,
but it is hard to move forward without it.

It is only after an experiment validates all or most of the hypotheses

underpinning its business plan that the question becomes “How do we
scale the new business?”

Scaling: Growing the New Venture

The issue of scaling or growing a successful new venture, while
always difficult, is less problematic in entrepreneurial firms where growth
is largely a function of attracting new capital and recruiting new people.
Scaling is also less of an issue for incremental innovations such as the
introduction of a new product or service where the new idea can be
integrated into existing structures and processes. But for a large firm,
scaling a new disruptive business or business model that moves from a
successful experiment to a fully operational business is a moment of both
comercial and organizational vulnerability. You have reached the point
where investment steps up a level. A well-designed and executed
incubation can improve the odds of success, but it cannot eliminate risk.
There is still a step into the unknown: Will customers behave and spend
money as indicated by the experiment?

There is also a temptation to push the accelerator too hard, too

quickly. “Never invest ahead of learning” is a golden rule for scaling.

Organizational vulnerability requires coordinating growth, but it is

also about managing internal political dynamics, an especially important
task with disruptive innovations. An explore unit that begins to scale
successfully becomes vulnerable in three ways.

First, if it is operated separately, it may be regarded by managers in

the core with skepticism, both because it is receiving resources that they
would like and may be seen as the “hot” new thing, relegating them to the
“old” business.

Second, many disruptive new businesses may initially offer lower

margins and cannibalize the exploit business, not typically appreciated by
leaders of the exploit business.

Third, if the new business begins to succeed in the market, there

may be a tendency to evaluate it with the rigors of the “exploit”
management system where it may be seen as underperforming financially.

To be successful at scaling, a new venture needs to add customers,

capacity, and capability fast enough to maximize the market opportunity.
What is particularly interesting is that in contrast to entrepreneurial firms,
large established firms typically have greater access to these assets and
capabilities and, done correctly, should be able to scale faster than new
companies. Just as with any business, it must do this in balance, so that it
does what it needs to do to achieve revenue growth without excessive
costs. The firms that are best at scaling appear to use all the following
options to meet the needs of the new venture; they are not dogmatic
about following a single approach:
 Acquire—Mergers and acquisitions (M&A) is an obvious means to
accelerate scaling for a new venture. If there are firms that can be
acquired that have the necessary access to customers, capacity, and
capability, then it can be na excellent fit.

Build—Another route is to commit to a significant investment and

build capacity internally.

Partner—A further option is to find partners with the resources that

a venture needs to scale.

Leverage—The greatest advantage an incumbent has over a startup

is that they start with potentially valuable assets and capabilities that the
startup lacks. The mature business has customers, capacity, and
capabilities, some of which can be repurposed to meet the needs of the
new venture.

The acquire and build options are often seen as the most attractive.

Acquisitions enable a firm to move at pace to assemble the

elements of a business and, as with the example above, capture the
opportunity quickly. However, the success rate of these sorts of
acquisitions is surprisingly low. Acquisitions that aim to “reinvent” a
business model substantially underperform relative to those that add to
existing capabilities. Many such acquisitions are also “startups” that may
have immature technology and unproven business models. This reality,
coupled with the difficulty of integrating startups into the corporate
culture, tends to drive the low success rate.

The challenge with adopting a build approach is that the likelihood

of corporate scrutiny and intervention increases. The result is often to
reduce the scope of the innovation so that it can be delivered with lower
risk in a shorter time. This pressure is particularly strong if there are new
technology assets involved that are outside the core competence of the
firm. It is challenging to stay on a long, uncertain path, sustaining
investment, with limited market data on the payoff.
 Partner and leverage offer a more complex approach. Both involve
reconciling competing interests and, often, navigating inter- or intra-
company politics. The payoffs, though, can be much greater.

There is no single right answer for scaling a new venture. Those that
are successful often combined two or more of these strategies.

Leverage is a successful, if underused, model for scaling new

ventures. What is essential is the role of leadership in providing a
supportive and enabling context. Three elements appear critical: active
sponsorship from senior leadership, separate explore and exploit units,
and ambidextrous leadership that can balance the competing demands of
explore and exploit.

Active Sponsorship from Senior Leaders—This is critical to both

getting new ventures the assets and capabilities they need and to enabling
them to override corporate norms and policies. Unless the new venture
has active seniorlevel sponsorship, the internal dynamics of the core
business are likely to slow down or smother the new business. Some of
this occurs because existing core units will demand compliance with their
processes (e.g., finance, HR). Some of it may come from a sense by the
exploit managers that the new venture is not important (a science project)
or, worse, it is directly competing with the core unit (e.g., for customers,
manufacturing, or talent). In most instances, these managers are making a
rational choice to do what, from their vantage point, will optimize
business performance. For explore units to succeed, senior leaders need
to be prepared to assert the logic of long-term priorities to ensure that the
requisite assets and capabilities are available to the new venture. That can
be uncomfortable for some leaders who may be conflict averse. If these
leaders have been involved as sponsors through the ideation and
incubation processes, then there is a higher probability that they will
understand why these battles are worth fighting.

Separate Explore and Exploit Units—No new venture can survive

without the autonomy to act with a faster clock speed than is typical for
the incumbente organization. Unless it has this autonomy, new ventures
will be forced into making compromises between what is needed to scale
the business and what is right for the core. This means having separate
architectures (people, structure, metrics, and culture) for the explore and
exploit businesses.

Ambidextrous Leadership—Scaling a new venture requires two very

diferente leadership competencies: the ability to lead the explore unit and
the ability to balance both the explore and the exploit businesses. In the
first instance, the explore leader needs to be entrepreneurial, able to
create a compelling vision, deal with organizational politics, recognize how
and when to pivot, and leverage the organization’s resources. In the
second instance, the leader needs to be able to deal with the tensions
inherent in running both an exploit and an explore unit, with different
time frames, skills, structures, metrics, and culture. This inevitably leads to
conflicts over resource allocation and priorities.

To solve the explore problem, one common strategy is to hire an

executive from outside the firm to bring expertise, knowledge, and
capabilities the firm lacks. Counterintuitively, putting an outsider in this
role is sometimes correlated with underperformance at scaling the new
unit. One outside leader hired to run an internal venture described how
the core business made no effort to actively oppose her efforts, but they
denied her any support and assistance. Her team was underequipped to
achieve its goals and could not get access to the resources of the core
business. She described how her team became isolated in its approach
—“we became the rebel alliance”—and totally failed to build the
relationships they needed to be successful. In contrast, internal
appointments have the credibility and social networks that enable them to
call in favors and leverage assets and capabilities that outsiders cannot.

Beyond appointing an appropriate leader for the explore venture,

there also needs to be a leader capable of managing the tension between
explore and exploit. Every successful story of scaling a new venture inside
an existing Corporation we are aware of involves a leader that can manage
this balance. They are able to drive operational performance in the
mature business, at the same time as the explore unit invests in building
the customers, capacity, and capabilities for the long term.
 The Three Stages of Disruptive Innovation in Action
Excelling at one or two of the three stages of disruptive innovation
is unlikely to lead to success. If a firm generates new ideas but cannot
adequately determine which are likely to succeed as a new business, they
will waste resources on unsuccessful ventures.

If a company is good at both ideation and incubation, they may

initially identify promising new ventures only to fail as they try to scale
them. Similarly, a company that has no skills at ideation but is good at
incubation and scaling is likely to grow businesses that are not truly
innovative.

Success requires all three disciplines.

AMAZON CASE: To operationalize these principles, Amazon has

made innovation a pervasive part of their culture. Their leadership
principles include not only customer obsession, but also invent and
simplify, think big, and a bias for action. To make innovation a part of daily
life, Bezos encourages small teams to continually come up with new ideas.
He says, “If you can increase the number of experiments you try from a
hundred to a thousand, you can dramatically increase the number of
innovations you produce.” To accomplish this, the emphasis is to
maximize the number of experiments but to keep the cost of each as low
as possible. He argues that the real barrier to innovation is not a lack of
imagination but the bureaucracy of large organizations.
 Idea Generation at AMAZON: The core process begins with a person
or team proposing a new idea that will enhance the internal or external
customer experience. To formally propose the idea, the originator begins
by composing a six-page narrative called the PR/FAQ (never a PowerPoint
presentation). The PR/FAQ follows a strict format. It is based on the idea
of “working backwards” from a customer problem and begins with a one-
page press release (PR) announcing the release of the final product. This
includes the name of the product in language that the customer will
understand, who the customer is, what benefits they will receive, a
description of what problem the product or service solves, and an
explanation for how the product will elegantly solve the problem. The PR
announcement may also indicate how easy it is to use the product and
even hypothetical customer quotes. The PR is followed by five pages of
Frequently Asked Questions (FAQs) that explain why customers will want
the product, how they will use it, how much it will cost, and what benefits
they will reap. It would also indicate the potential market size and any
risks associated with the product. The FAQ is often supplemented by an
appendix and a mock-up of a prototype.

Once the narrative is prepared, it is presented to a group consisting

of colleagues, the immediate boss, someone from higher management,
and a “bar raiser” who is from another function who can act as a sanity
check. Again, there is a strict process to be followed. The narrative is not
circulated in advance but passed out in the meeting, and the first 30
minutes are in silence as participants read the memo. This format ensures
that all those in the meeting begin with a common understanding of the
proposal. In evaluating the proposal, the discussion focuses not just on the
technical merits of the idea, but also whether it has the potential to be a
big idea that the customers would really love. After a discussion, the most
senior person in the room will make a decision to allocate minimal
resources to continue the project. These may include release time for the
proposer, a small budget, or some engineering support. Estimates are that
about 50% of proposals move to the next stage. If successful, the revised
PR/FAQ will then be presented at the next higher level of management.

Several aspects of this process are noteworthy with respect to

ideation. First, like design thinking, it is a bottom-up process that begins
with an obsessive focus on improving the customer experience. Second,
the structure of the meeting also acts to focus the discussion on an
evaluation of the specifics of the proposal with everyone having the same
information. Finally, the Amazon culture is one that encourages people at
all levels to identify and propose incremental innovations to enhance unit
productivity and drive efficiency or new business ideas that leverage
existing assets and capabilities. Estimates are that the business leaders at
Amazon will see about 100 PR/FAQs a year, suggesting that the process is
successfully generating a constant stream of new ideas.

Incubation at AMAZON: Once approved and with some limited

resources, the next step is to build a MVP and to quickly get this in front of
customers. Like the lean startup methodology, the priority is not on
internal testing, but getting the product launched and learning quickly
which features are valuable and which are not. This is usually done using a
small market niche. Customers are probed as to whether they would
actually buy the product. Do the early adopters love it? Can it lead to a big
business? Like the business canvas model, variants of the product are
often explored using A/B tests. The concern in this phase is not
profitability or the competition, but whether the product or service is
strategically importante and will really deliver a better customer
experience. Sometimes, teams employ the business canvas to ensure
completeness.

To enhance speed and accountability during incubation, Amazon

relies on what are called “single-threaded teams” or “2-pizza teams.”
These are small teams (that can be fed with two pizzas) of six to ten
people. Each team consists of a small number of people with the skills
needed to develop the idea (e.g., a handful of engineers, a product
manager, and a designer). The team is free to act autonomously with little
or no need to coordinate across functions. There is often a single agreed-
upon metric used to provide focus and accountability. Unlike many
organizations where internal projects are required to use a standardized
technology or set of tools (like Google), teams at Amazon are able to use
whatever tools and technologies they think are best for the task at hand.
This single or small set of feedforward metrics and the freedom to use
whatever approaches are most useful becomes the equivalent of a P&L so
that the team itself becomes like a small profit-and-loss center, and the
team leader is like a mini-general manager. Often, the person who
proposed the idea becomes the team leader, so that these projects act to
attract and retain entrepreneurial talent.

The use of small, highly focused, and accountable teams, the

emphasis on an MVP, fast iteration, and a customer-centric focus replicate
the lean startup methodology but do this in a large organization with
access to more talent and resources than a startup. The use of small
teams also enhances the speed with which experiments can be run.
Experiments are designed to fail early. To facilitate this, a distinction is
made between one-way and two-way doors. With two-way doors, the
consequences of a failure are minimized because you simply return to
where you started. Because the teams are self-contained, lengthy and
costly coordination across functions is minimized. There is an
acknowledgment that this approach may result in some duplication and
inefficiency, but the benefits of rapid iteration and learning outweigh
these costs.

If the product or process proves viable, the team may modify their
PR/FAQ and submit the next request to more senior management. This
request includes an account of the resources needed to begin to scale the
program. If approved, the team will be given the additional resources to
begin to roll out the product or servisse on a larger scale.

Scaling at AMAZON: The use of small decoupled teams also makes

scaling easier. As they begin to grow, teams (especially in product and
engineering) continue to own the product or service on an end-to-end
basis. They interact with other parts of the organization in a manner
similar to APIs (application programming interfaces). This permits teams
to leverage assets and capabilities of the larger organization (e.g., access
to capabilities, capacity, and customers). They request specific inputs and
outputs but continue to have their own budget, so maintaining control of
the project. This reduces dependence on others and the negative effect of
gatekeepers. With senior management oversight, additional resources are
provided to ensure that the project is receiving what is needed to scale
the effort. A key element of this is the widely shared belief that to be truly
innovative, you need to be prepared to be misunderstood for a long
period of time and willing to persist in the face of skepticism.

Supporting this process is a tenet referred to as “the institutional

yes.” In most organizations, approval for additional resources is made
through a committee or review process in which a single veto can either
kill or slow down an initiative. Managers at Amazon recognize this, and
rather than ask “why are we doing this?” ask “why not?” They believe
that most big errors are not of commission but omission. For example,
Amazon Go, the new retail store with no checkout, was begun by
envisioning how great it would be for customers to simply walk in, grab
their item, and walk out without having to wait in line to pay. Presented
with this idea, most organizations would say “Why do we need this?” and
note that it required skills that currently do not exist (sophisticated
cameras and artificial intelligence). Amazon said “Why not?” Once there is
evidence of early success, the program does not compete for resources
with others but is funded on its merits. The emphasis is to double down
quickly on winners and not get caught up in trying to coordinate across
projects.

It is worth noting that it is the combination of ideation (the PR/FAQ

process), incubation (single-threaded teams with MVPs), and scaling
(institutional “yes” and the escalation of resources with senior manager
oversight) that accounts for Amazon’s remarkable success at innovation.
Many of these projects begin as incremental improvements but morph
over time into new businesses. It is this process that underlies Amazon’s
ability to move into new businesses in cloud services, third-party
fulfillment, logistics, retail sales, and consumer technology.

IBM Emerging Business Opportunities

IBM company implemented a process to discover and develop new
businesses from within the larger organization. Their EBO process
developed a set of businesses that generated more than $5 billion in
growth over that period. This process began with a clear strategic intent
that included a desire to explore new business models and capabilities,
generate $1 billion in revenue over a five-year period, offer sustained
profit growth, and establish IBM as a market leader in these new areas. It
recognized the need for separate operating units and ambidextrous
leadership and replicates the ideation, incubation, and scaling disciplines.

Ideation at IBM: The EBO process begins with the appointment of a

very senior executive as the person responsible for overseeing the new
ventures. Think of this person as the internal venture capitalist who will
fund, oversee, sponsor, and, if necessary, terminate the new venture. The
process began with a set of principles for deciding on what types of new
growth businesses IBM wanted. They identified six characteristics:

••The EBO should be aligned with and support the larger IBM
strategy.

•• New ventures should provide for cross-IBM leverage (hardware,

software, and consulting).

•• It should offer a new source of customer value.

•• It should promise revenues of $1 billion within a five-year time

frame.

•• It should allow IBM to be the market leader.

•• It should provide sustained profit and not be commoditized.

With these constraints, twice a year, the company would survey

employees, IBM fellows, technology leaders, and venture capitalists. Given
their insights into IBM, they were asked for suggestions for opportunities
or areas in which IBM might grow new businesses. Every year, they would
receive on the order of 150 suggestions. Experts in the strategy group
would then review these suggestions and narrow down the list to 20 to 30
areas that they believe might justify some investment. Strategy teams
would then collect market insight data on these áreas and further narrow
the list to a small set of potential new businesses. Based on the input of
senior management, they would then focus on selecting leaders for these
initiatives and provide funding for them.

Incubation at IBM: Once an idea for a new business is selected and

a leader appointed, a business plan is developed that includes an initial
allocation of resources (money, people, and technology) and a set of
milestones and metrics that will determine progress and subsequent
funding. A training program for the leaders helps them understand how to
establish and communicate a clear vision, select a team, understand the
organizational politics, and sustain an initially unprofitable venture. The
leader of the new business reports both to the senior management of a
line of business and to the senior corporate executive responsible for new
ventures. The EBO is then run as an internal startup with continued
funding based on meeting set milestones just as a venture capitalista
might fund a venture.

Scaling at IBM: What differentiates the EBO process from the lean
startup methodology is the careful attention that is paid to growing the
new venture. To ensure that resources are provided in line with growth,
the new venture relies on the oversight and support of the senior
corporate leader, disciplined mechanisms for cross-company alignment,
and resources that are ring-fenced to make sure that funding is provided
when and where it is needed. Growth is closely monitored and, if
milestones are not met, the initiative is stopped. If milestones are met,
resources continue to flow to the new venture, any resistance from other
parts of the organization is moderated, and a clear process is in place to
gradually migrate the new venture back into the mature business. It is
only when the EBO has a strong leadership team in place, a proven
customer value proposition, and clear market success that the new
business is migrated back into the larger organization.

Like the Amazon example, the IBM EBO process illustrates how
careful attention to all three innovation disciplines is needed to organically
grow new businesses. Other firms such as Bosch (the German industrial
company) and AGC (the Japanese materials firm) have developed similar
programs that pay careful attention to ideation, incubation, and scaling.
For these efforts to succeed, what is important is that the process
encompasses all three disciplines.

Conclusion
As the threat of disruption has increased, academics and
practitioners have increasingly focused on how organizations can
innovate. From the practitioner side, great progress has been made in
helping firms with ideation and incubation. Processes such as design
thinking, open innovation, internal innovation programs, and the lean
startup methodology have been successfully applied.
 However, these programs have most often been used to help firms
increase incremental innovation and have proven to be less useful for
helping them meet the challenges of disruptive change. Furthermore,
because these approaches were designed originally for entrepreneurial
firms and not incumbent corporations, they have largely failed to solve the
scaling issue for large firms.

To deal with discontinuous innovation, firms must master three

distinct stages or disciplines, idea generation (ideation), incubation
(validation), and growth (scaling). While these three stages can also apply
to incremental innovation, being successful at discontinuous innovation
requires leaders to be more sophisticated at managing these stages (see
Table 1), especially in how they approach scaling.

Furthermore, mastering only one or two of these stages is

insufficient. Having new ideas that do not meet the market test, having
market-tested ideas that cannot be scaled, or scaling ideas that are not
market validated are all recipes for failure. Success needs all three.

For large firms, simply “acting like a startup” is not enough to

guarantee success. The organizational and cultural inhibitors of success
inside a Corporation remain formidable. Our fundamental proposition is
that for new ideas—even those with demonstrated market acceptance—
to become new businesses, they require leaders to master all three
disciplines. This can be achieved through a blend of acquisition, building,
partnering, and leveraging assets and capabilities from the exploit
business. To do this successfully requires leadership and organizational
practices that manage the tensions of successful incumbent organizations,
which the startup-inspired methodologies were not designed to address.
Each of the three stages—ideate, incubate, and scale—is distinct and
necessary, but only when all three are in place is it likely that new ideas
will result in new business that enable incumbents to lead disruptive
innovation in their markets.
 Exploitation Exploration
Ideation  Open Innovation (getting ideas Same but with:
from others outside the firm)  Broader scale of
 Corporate Venture Capital ambition
(getting ideas from the startup  Defined hunting zones
world)
 Design Thinking (getting ideas
from customers/users)
 Employee Involvement
(getting ideas from employees)
Incubation  Lean Startup (build-measure- Same but with:
learn to work backwards from a  Hypothesis Testing
business goal) (development of
 Business Canvas (design and capabilities through
test a new venture hypothesis) experimentation)
 Launch Pad (identify target  Feedforward
customer pain points, interviews, Measurement (metrics
iterate to achieve product-market of progress toward goals)
fit)  Governance (leadership
to ensure assets and
capabilities are available as
needed)
Scaling  Spreading Constructive Ideas Same but with attention
(going slower to get faster, reduce to:
cognitive load, more vs. better,  Customers, capacity,
breach assumptions, promote
capabilities obtained
accountability, clearing the path)
through acquisitions,
building, partnering,
and leverage
All with:
 Active senior sponsorship.
 Separate explore and
exploit units.
 Ambidextrous leadership.
BEST ARTICLE AWARD 2020: The Three Stages of Disruptive Innovation: Idea
Generation, Incubation, and Scaling. Charles O’Reilly and Andrew J.M. Binns (©2019
The Regents of the University of California)