The four-stage innovation framework Manifold uses to quickly navigate complex problem spaces and arrive at novel solutions.
Manifold uses a specific definition of innovation that we believe provides utility for innovators. But our definition is still pretty broad. Given the work we do, certain types of innovation are more meaningful than others.
We propose a framework for distinguishing innovation based on four levels of increasing complexity. This approach is neither the only framework — there are many — nor necessarily the best for every purpose. But we do believe that it is the best framework for understanding how to pursue different types of innovation most effectively.
Consultants and academics often codify their experience into frameworks, designed to abstract experience into principles that can be broadly applied to a variety of situations. By operating within frameworks, they’re able to quickly navigate complex problems and arrive at novel solutions.
Many experts have described ways to distinguish between different types of innovation. What vectors of differentiation do they tend to use?
Our goal is to identify a consistent way to differentiate between various types of innovation that will inform us about when it’s appropriate to change our process and decision-making.
Some distinguish innovation based on the extent of novelty, complexity, or change involved. The OECD describes four different types of innovation, one of which is incremental innovation:
"Incremental innovation: innovation that builds closely on technological antecedents and does not involve much technological improvement upon them."
“Incremental” and “sustaining” are two words commonly used for modest or low risk innovation. In a corporate context we often talk of “exploitation” (versus exploration) as well. The other end of the spectrum is described with a variety of words, each of which incorporates some nuances or variations of meaning. Examples include:
Disruptive innovation is a term coined by Clayton Christensen in the mid 1990’s to describe a specific form of creative destruction which is rooted in Joseph Shumpeter’s concept of creative destruction.
Steve Blank called out Christensen (and Shumpeter’s) theories in a 2014 speech:
"Disruptive innovation – this is the innovation we associate with startups. This type of innovation creates new products or new services that did not exist before. It’s the automobile in the 1910’s, radio in the 1920’s, television in the 1950’s, the integrated circuit in the 1960’s, the fax machine in the 1970’s, personal computers in the 1980’s, the Internet in the 1990’s, and the Smartphone, human genome sequencing, and even fracking in this decade. These innovations are exactly what Schumpeter and Christensen were talking about. They create new industries and destroy existing ones. And interestingly, in spite of all their resources, large companies are responsible for very, very few disruptive innovations."
Specifically, Christensen’s term described innovation by upstarts targeting unmet needs in “low-end footholds” that the incumbents had ignored or been unaware of. He also clarified that for something to be disruptive (according to his definition at least), the existing market had to consider it undesirable at first. He went on to argue based on these tests that Uber is not a disruptor. Commence head scratching.
It appears, by the way, that Steve Blank probably understood disruption to be a broader concept than Christensen intended. After all, automobiles, television, and the fax machine are hard to fit into the notion of a low-end foothold.Christensen has since walked back from his position that Uber is not disruptive, but the fundamental fact remains that the term he popularized—to him at least—is quite narrow, and rooted primarily in Joseph Shumpeter’s original concept of creative destruction.
The concept of creative destruction, while useful, is still largely a theoretical construct. Christensen inarguably contributed to economic theory with his notion of disruptive innovation, but it’s not clear to us due to its narrowness that it offers a particularly useful rubric for understanding innovation.
When most people use the word disruptive innovation, they probably mean the broader notion of innovation that disrupts an industry. That’s the meaning we assign to it at Manifold unless we specifically invoke Christensen.
In this broader conceptual vein, using the extent of innovation to distinguish types of innovation seems useful. If you’re doing something incremental, it seems more likely that traditional management orthodoxy would be appropriate. If you’re doing something disruptive, it seems much more likely that you would need to adopt new strategies to accommodate. But still, the terms we’re discussing feel like blunt instruments that don’t offer enough utility for our purposes. Let’s keep looking.
Another way to think of innovation involves what you’re trying to improve. For example, the same OECD report mentioned above describes product, process, marketing, and organizational innovation. Each of these appears to distinguish between the types of resources or capabilities an organization is trying to improve. An example:
"Product innovation: development of new products representing discrete improvements over existing ones."
A different approach has been written about by Doblin in its Ten Types of Innovation. They focus on the ten different innovation targets, including: profit model, network, structure, process, product performance, product system, service, channel, brand, and customer engagement.
It’s definitely a worthwhile read. Jay Doblin, the cofounder, was an iconic designer with expertise in systems thinking and design theory. As a result, their approach tends to be derived from a product design perspective, and may not be as useful when it comes to the broader context, and specifically in the context of technology and business model innovation.
Steve Blank also talks about process innovation, but his version is different and would probably be labeled incremental or sustaining innovation by most:
“Car companies introduce new models each year, running shoes grow ever lighter and more flexible, Coca-Cola offers a new version of Coke. Smart companies are always looking to make their current products better – and there are many ways to do this. For example they can reduce component cost, introduce a line extension or create new versions of the existing product. These innovations do not require change in a company’s existing business model.”
In fact, today this sort of innovation seems a lot more like everyday execution than innovation.
Steve also talked about business model innovation, which probably falls under the “what you’re trying to improve” umbrella.What’s a business model? That has been described many different ways, much like the definition of innovation. And similar to innovation, we have a preferred definition, specifically Blank's:
“A business model describes how your organization creates, delivers, and captures value.”
Business model innovation and disruption often go hand-in-hand; successfully creating new ways of creating, delivering, and capturing value is often disruptive. (By the way, around Manifold you’ll hear us use the term “value creation model” frequently in place of business model. That’s a nod to the fact that commercial value creation isn’t the only game in town.)
Is defining innovation based on what you’re trying to improve the best approach for our purposes? Probably not.
Innovating a product or process, for example, probably has a lot of similarities from the perspective of team, organization, and the path it takes—and as a result aren’t as useful for our clients trying to take a strategic approach to innovation.
It’s probably more important to understand how radical the change is than what’s being changed. The exception might be business model innovation due to its likely implication of breakthrough or disruptive innovation (when successful).
Ralph-Christian outlines a concept he calls “Dual Corporate Innovation Management” that involves a balance between exploitation-oriented and exploration-oriented initiatives.
He identifies three (time) horizons for innovation based on balancing exploitation (“enhancing existing business models or technological capabilities”) and exploration (“novel, and often disruptive, business models or technological capabilities”) for establishing organizations. He advocates an innovation management system based on a portfolio approach driven by a gap analysis. Importantly, he points out that exploitation and exploration require different approaches to achieve success. In the end, he describes three horizons of innovation: core, growth, and future. His approach is useful from a corporate strategy perspective, and is rooted in traditional management orthodoxy. But it’s not really aimed at distinguishing different types of innovation.
McKinsey also describes three horizons of growth:
As they say, “Time, as noted on the x-axis, should not be interpreted as a prompt for when to pay attention—now, later, or much later. Companies must manage businesses along all three horizons concurrently.”
This seems like a sort of GTD (Getting Things Done) for corporations (Inbox, Today, Someday). It’s useful as a tool for allocating time and energy, but not as useful for understanding the nature of the innovation you’re undertaking, and not particularly informative about how you organize, lead, and manage the team that’s involved. It’s also not particularly useful outside of the context of an existing corporation.
Greg Satell describes a matrix-style approach he describes as the “4 Types of Innovation.” His model incorporates the notions of domain and problem, asking the following questions:
From that, he outlines a four-part matrix that includes:
A key insight he offers is that when the marketplace shifts, often “innovating your products won’t help—you have to innovate your business model.” We find his approach to be insightful and interesting. But we believe his emphasis on problem as one of the vectors misses an opportunity. Whereas clarifying a problem is very important, particularly early on, we don’t believe it’s as useful to focus on that vector.
Manifold borrows from Greg Satell’s approach, but replaces the system (e.g., the value creation model) for the problem. Our two vectors:
We also change the nature of the questions asked. Instead of asking how well the vectors are defined, we ask “how familiar is it?” The answer ranges along a spectrum from familiar (to the team), to unfamiliar (to the team), to novel (to the world):
Known to me — New to me — New to the world
This enables our framework to view the vectors more simply and objectively. It also more explicitly accounts for both the novelty and the relationship between the novelty and the organization involved.We believe this new approach changes the model significantly, making the framework more useful. For example, this model makes it more natural to recognize tiered levels of innovation, which we found to be useful when we first looked at distinguishing innovation based on extent (above). It also recognizes that innovation can be experienced differently for different organizations.
Manifold's four different levels of innovation:
This is a known system of value creation using known resources and technology. These sorts of innovation tend to have clear proxies, and offer predictable timing, budget, and resource requirements. It’s easy to measure and manage risk, and it’s reasonable to expect a steady upward progress. It’s arguable this isn’t even really innovation, although we think it’s fair to use the term here as long as we’re careful to differentiate from other types of innovation.
There’s nothing truly novel (to the world) here, but something is unfamiliar to the team. Perhaps it’s the implementation of new (to you) technology or a new (to you) revenue model. This sort of innovation still offers a relatively predictable path, but it’s riskier and harder to predict than level 1. These innovations tend to take longer than level 2 innovations.
Level 3 innovation often involves new technologies and business models, albeit not radically so. Usually there are somewhat close proxies or partially understood or known technology. It starts to become hard to predict path, timing, budget, and required resources. Risks are very hard to assess. Often the complexity is in the combination. This sort of innovation tends to take a longer time.
Breakthrough innovation is the process of bringing to market a truly novel business model or domain. Either or both of the system or domain is new to the world. These sorts of innovations tend to take a very long time and are fraught with risk and uncertainty. They also tend to create significant value when they’re successful.
This approach to distinguishing types of innovation has the benefit of simplicity, with only two factors to consider. Each of the levels also has natural implications for key factors such as level of uncertainty and risk, volatility, predictability, typical timelines, etc. That’s useful because these are the factors that typically force changes in strategy, organization, and process to maximize the odds of success. By tying them naturally to the levels of innovation, we are increasing the usefulness of the framework.
The framework also works both in the context of a corporation, and for startups. And it recognizes that what might be transformative for one organization might be mostly change or optimization for another.It has some weaknesses as well. The rating of the vectors is subjective. The model also may miss nuances (or perhaps substance) in the differences between innovation based on system vs. domain. After all, core technology innovation likely has meaningful differences from business model innovation.
Overall, however, we have found it a useful approach to differentiating innovation for the purposes of understanding how best to drive efficient value creation for our clients.