Can innovations be classified at all, and are there genuinely several types of them? The question sounds academic and it is the difference between a team that knows what it is doing and a team chasing whatever is fashionable this year.
I went looking for an answer because I needed one, not because the subject interested me in the abstract. What I found was a field that has never agreed on how many types exist, and that turns out to be the wrong thing to be arguing about.
The published frameworks disagree because they are classifying different objects and almost none of them says which. One sorts fields of application, another sorts people, a third sorts stages of a process, and all three call the result a type of innovation.
So the useful question in any innovation classification is not how many types there are. It is what is being classified, and then whether the answer tells a specific person what to do on Monday.
Which is what a vehicle programme called Renovatio taught me, several years before I had the vocabulary to describe it. Before the definitions are useful, it helps to be clear on what innovation and invention actually mean, which I have set out in the four terms most organisations use as one.
| The apparent problem | The actual problem | What a usable answer does |
|---|---|---|
| Nobody agrees how many types there arePublished frameworks offer three, four, ten or more, and every innovation classification contradicts the ones beside it. The field reads as unsettled, which sends most organisations back to the fashionable term of the moment. | They classify different objectsOne framework sorts fields of application, another sorts personality profiles, a third sorts stages of an ideation process, and each calls its categories types of innovation. The disagreement is about the object being classified rather than about the count. | It assigns a role, not a labelA classification is useful when a specific person can read it and know which kind of work suits them, which turns innovation into each person’s business rather than everyone’s business. Everyone’s business is nobody’s. |
Why I needed a classification at all
This started with a programme rather than with a theory. Renovatio was an attempt to reinvent how vehicles were built and sold, and I wanted the whole team involved in designing and making it.
The programme and what I wanted from it
The intention was straightforward and, I now think, naive. Everybody would feel involved in the conception and the construction, and the work would gather the team around a shared value of questioning things.
I had a rough acquaintance with Clayton Christensen’s work at the time, and I had been using design thinking techniques since the early 2000s. Some of those went into the programme, including making modifications to client vehicles first so we could hear what people thought before anything went into ours.
Several things then became apparent, and none of them was in any book I owned.
Four ways the team turned out to differ
The first difference was which part of the work interested them. Some wanted to shape the concept and had no wish to go further, arriving with things like putting the R of Renovatio on the front. Others were already picturing themselves testing it, and knew a stretch of track beside the railway line with a useful bump for checking the suspension.
The second was their attitude to the project itself. Some could not see the point of developing our own vehicle when we were already working on client vehicles that performed perfectly well. Others immediately offered to come in at the weekend to help.
The third was how many ideas they produced. A few had an uninterrupted flow and arrived every morning with a fresh harvest of varying feasibility. Others had clearly thought about it already and were far more pragmatic, arriving with a single view such as it would need at least four hundred horsepower.
The fourth was how bold the ideas were. They ran from entirely expected to completely unhinged. Some framed their thinking against vehicles they knew, wanting a chassis fifteen centimetres wider than the reference model. Others wanted to start from nothing, asking what if we offered no options at all, or what if the updates were free.
What that made obvious
Four things bothered me at the time and they add up to one problem. Running the sessions properly required knowing the techniques and having somebody to facilitate them.
The word innovation itself seemed vague and frankly fashionable, with every expert defending their own territory and nobody agreeing on terms. Christensen, Jean-Marie Dru and Tom Peters were all using disruption to mean different things.
Digging further left me buried in terms that did not fit together: continuous improvement, aggregative innovation, radical innovation, disruptive innovation, breakthrough innovation and cross-pollination, the last of which I had first heard in Australia while commercialising a sports-agent approach for senior engineers.
And the level of abstraction in the books, from Schumpeter through to Tim Brown, was too detached from the ground for colleagues who handled a spanner considerably better than a concept. In summary, it was a mess.
Which produced the thought that carried everything afterwards. There must be several levels of innovation corresponding to several kinds of profile, matched to my culture and my project, which would let each person be involved according to their role rather than in general.
The problem with the published classifications
Once I went looking properly I found an elegant example of complexity that left me none the wiser. In France or anywhere else, nobody agreed on which types of innovation exist or how many of them there are.
They are not classifying the same object
It took me an embarrassing amount of reading to notice why. The frameworks contradict each other because each one is sorting a different thing, and the word type is doing four jobs at once.
Larry Keeley’s ten types include service, brand and network while omitting management, and it appears to rest partly on the four types the OECD proposes in its work on defining innovation, being product, process, marketing and organisation. In my view it confuses a field of application with a type of innovation, which are different questions.
Tom Kelley’s ten faces describes ten profiles of innovative people, with examples that have worn thin. Those profiles have nothing to do with Keeley’s ten types, and they mostly concern continuous innovation and design thinking techniques, which is the business his firm is in.
Gijs van Wulfen’s innovation maze considers the ideation process from the idea through to the business challenge, which is a sequence rather than a taxonomy. And in 1990 Rebecca Henderson and Kim Clark published their paper on architectural innovation, which describes several versions of continuous innovation rather than several types.
Four frameworks, four different objects
Set them side by side and the pattern is immediate. Read the object column rather than the count column, because the count is what everybody argues about and the object is what actually differs.
| Framework | How many | What it actually classifies | Useful for |
|---|---|---|---|
| OECD definitions | Four | Fields of application | Statistical comparison between countries |
| Keeley, ten types | Ten | Fields of application, expanded | Auditing where a business has and has not innovated |
| Kelley, ten faces | Ten | People and their roles in a session | Composing a design thinking team |
| Van Wulfen, innovation maze | A sequence | Stages of an ideation process | Running a project from idea to business case |
| Henderson and Clark | Four | Degrees of technological and architectural change | Explaining why established firms fail |
Every one of those is a good answer to the question it is actually asking. None of them answers the question a leadership team asks, which is what kind of innovation we should be attempting and who should be doing it.
The list does not stop there, and neither do the books proposing stages, styles, types, essentials or facets.
The three that arrive somewhere similar
Three classifications published since then land close enough to my own to be worth examining, including one from the author whose work started the argument.
Where Christensen’s own three categories land
Christensen identified three categories in his Harvard Business Review article on the capitalist’s dilemma: performance improving, efficiency, and market creating.
My reading is that this goes in a circle. Performance improving is continuous innovation under another name. Efficiency and market creating both fall, depending on the case, into either disruptive or breakthrough work.
That is a compliment rather than a criticism. A framework built for a different purpose arriving at overlapping categories suggests the underlying distinctions are real, and the two traditions the word disruption comes from are set out in the comparison of disruptive and breakthrough innovation.
Nielson, Satell and the convergence
Jake Nielson published four types on his Ignition Framework blog, including an integrative category. What I like about his approach is that it stays concrete and looks at the financial results of innovation rather than only at the concept of it.
Greg Satell reached four types in his Harvard Business Review piece on the four types of innovation and the problems they solve: basic research, sustaining, disruptive and breakthrough.
His basic research maps closely onto what I call adjacent, since that type of work amounts either to research and development or to drawing on sources outside the business, meaning conferences, journals and researchers.
Two of Satell’s four share names with two of mine, which is worth stating plainly rather than leaving a reader to wonder. Independent arrival at similar distinctions is the outcome you would expect if the distinctions describe something real.
Field note
The people who opt out in the first ten minutes
This is a pattern across workshops rather than one client story, and I would rather label it honestly than present a composite as a case study.
Ask a room who the creative people are and something consistent happens. Three or four heads turn towards the same two individuals, those two look mildly embarrassed, and everybody else has left the exercise before it started. They are not being modest. They have made an accurate assessment that whatever is about to happen is not aimed at them, and in a session run on the premise that innovation is everyone’s business they are correct, because a demand addressed to everybody is addressed to nobody in particular. The interesting part comes later. Split the same work into four distinct kinds and ask people to choose one, and the people who opted out claim a category within a minute, usually the one closest to what they already do well.
The lesson I take into every session now is that the classification exists to let people opt in rather than to sort the work for an analyst. A framework that produces a tidy taxonomy and no volunteers has failed at the only job that matters in a room. Innovation is each person’s business, and the difference between that and everyone’s business is the difference between a room with four working groups and a room with two tired volunteers.
What a usable classification has to do
All of which produces a short specification. A classification worth having states what it is classifying, and then tells a named person what part of the work belongs to them.
Declare the object before the categories
The first requirement costs nothing and almost nobody meets it. Say whether you are sorting fields of application, degrees of change, stages of a process, people, or kinds of strategic bet, because those are five different exercises.
Mine sorts strategic bets, meaning the kind of move a business is making and what it requires to survive. That is why it cannot be compared directly with the OECD’s fields or with Kelley’s profiles, and why arguing about whether there are four types or ten is a category error rather than a disagreement.
Twenty years of collecting these books produced one genuinely useful shortcut, which is that the colour of the cover predicts the subject with unnerving accuracy. Blue for change and transformation, white for reference guides, orange for anything radical, red for creativity. That observation says more about the state of the field than my critique does.
Each person’s business, not everyone’s business
The second requirement is that somebody can read it and know what belongs to them. Which lets an organisation stop repeating that innovation is everyone’s business, a phrase the quality profession has been using since the 1980s with the same result.
Four roles cover most organisations, and each one owns a different question.
| Role | What they own | The question they answer |
|---|---|---|
| Leadership | Strategy | Which kind of innovation are we attempting, and can we carry it? |
| Human resources | Culture | What do we reward, promote and tolerate, and does that support the attempt? |
| Management | Managerial style | How do ideas get heard, and what happens to the ones we do not use? |
| Team member | Creative profile | Which kind of this work suits me, and where do I want to be involved? |
The nuance between each person’s business and everyone’s business matters more than it looks. One assigns a question to a named role. The other assigns responsibility to a crowd, and a crowd declines it.
Worth saying plainly that the fourth row is provisional rather than permanent. Renovatio showed people sorting themselves by what interested them on that project, not by what they permanently are, and a classification that hardens into a personality label has become the thing it was built to replace.
Want the classification itself?
Convinced the published frameworks answer a different question? Read the four innovation strategies and how to run a working group for each, which is the classification this article argued its way towards.
How I can help you sort this out
Nobody I work with needs another taxonomy. They need somebody to establish which question their organisation is actually asking, and then to hand each person in the room something that belongs to them.
Workshops and creative profiles
The workshop runs the sorting exercise from the field note, live, which is faster than explaining it and considerably more convincing. Most of the value is in watching the people who opted out claim a category.
The creative profile work goes deeper on the fourth row of that table, and the collaboration and design thinking material supports the first three. There is more on both in the creative profile tool.
Keynotes and diagnostics
Keynotes suit the moment when a leadership team has read four frameworks and adopted none of them. Explaining that those frameworks answer different questions removes the paralysis without discrediting anybody’s reading.
The diagnostic establishes which of the four roles in that table is currently unowned, which is usually the second one, and what that absence is costing.
If you want to test this, a conversation beats a proposal. You can tell me which framework your organisation has adopted and we can work out what question it was built to answer.
Conclusion: ask what is being classified
The published frameworks are not in conflict. They are answering different questions while using the same word, which is why comparing their counts produces an argument nobody can win.
So the first move in any innovation classification is to declare the object. Fields of application, degrees of change, stages of a process, people, or strategic bets, and each of those has a legitimate framework already built for it.
The second move is the one that decides whether anybody uses the result. A classification that lets a named person recognise their own part of the work turns innovation into each person’s business, and one that produces a tidy diagram leaves it as everyone’s business, which is nobody’s.
Renovatio taught me that before I had the words for it. Four people wanted four different things from the same project, and no book on my shelf, whatever colour its cover, had a category for any of them.
Frequently asked questions about innovation classification
Why do innovation classifications contradict each other?
Because they classify different objects while using the same word. The OECD and Keeley sort fields of application, Kelley sorts people, van Wulfen sorts stages of a process, and Henderson and Clark sort degrees of technological change. Their counts differ because their subjects do.
How many types of innovation are there?
The question has no answer until you say what is being classified. Published frameworks offer three, four, ten or a sequence, and each is a reasonable answer to its own question. Ask instead what you need the classification to decide, then choose the framework built for that.
What is wrong with saying innovation is everyone’s business?
A responsibility addressed to everybody is accepted by nobody in particular. The quality profession has used the phrase since the 1980s with the same result. Assigning a specific question to leadership, human resources, management and each team member makes it each person’s business, which people do accept.
Is classifying people by innovation style just a personality test?
It becomes one if the label sticks. The assignment should be provisional and per project, since people sort themselves by what interests them in a particular piece of work rather than by a fixed disposition. A classification that hardens into a permanent label has become the problem it was meant to solve.
