What Is Innovation?

Eight Recursive Fields from Perception to Use

A company holds an innovation workshop. Several teams develop ideas, cover walls with sticky notes, select their favourites and build initial prototypes. The presentation to senior management goes well. Six months later, hardly any trace of the results can be found in day-to-day operations.

Did the innovation fail during implementation? Perhaps. But the problem may have arisen considerably earlier. Perhaps the wrong observations were chosen as the starting point. Perhaps the question was framed too narrowly. Perhaps relevant knowledge was missing. Or perhaps an unusual idea was expressed but not recognised as valuable.

What is innovation if an idea alone is evidently not enough?

Innovation is neither a single flash of inspiration nor a linear sequence extending from idea generation to market launch. It emerges through a recursive social and organisational process: organisations perceive signals, construct problems, connect knowledge, develop possibilities, process different perspectives, make decisions, generate new insights through experimentation and translate viable new approaches into actual use.

The eight fields presented in this article are not an established standard model in the research literature. They form a research-based synthesis of different strands of innovation, creativity, team and design research. Their purpose is not to divide innovation into eight tidy phases. Rather, they make visible the points at which organisations gain or lose insight.

What Is Innovation, and When Is an Idea Still Not One?

The OECD and Eurostat’s Oslo Manual defines an innovation as a new or improved product or process that differs significantly from what existed before and has either been made available to potential users or brought into actual use within an organisation. The decisive factor is therefore not novelty alone, but introduction or use.

An observation is not yet an innovation. Nor is an idea. Even a functioning prototype may initially be no more than the outcome of an innovation activity.

This distinction matters to organisations because creativity and innovation involve different requirements. In organisational research, creativity generally refers to the development of ideas that are both novel and potentially useful. Innovation additionally includes the promotion, further development and implementation of those ideas. A company can therefore be exceptionally rich in ideas while possessing little capacity for innovation.

The Innovation Process Is Not an Assembly Line

1. Perceiving Signals and Deviations

Innovation often begins with an irritation: customers bypass an intended process. A team develops unofficial workarounds. A performance indicator improves while complaints increase. A new technology makes something possible for which no articulated demand yet exists.

Signals do not, however, speak for themselves. What is considered relevant depends on where an organisation directs its attention, what knowledge is available and which explanations are regarded as plausible. Metrics capture only what has previously been operationalised. Escalation routes transmit only what people consider worth reporting and possible to articulate.

Deviations in particular are often normalised too quickly: as an isolated case, resistance, insufficient competence or an individual error. A potentially revealing irritation is thereby converted into a matter that can be closed administratively.

A useful organisational question is therefore: Which deviations do we investigate, and which do we explain away so quickly that we no longer have to learn from them?

Later fields can reopen perception. An experiment produces new data. A different perspective reveals a previously invisible pattern. Perception is therefore not merely the beginning of the process; it remains valuable throughout it.

2. Constructing and Framing Problems: How Innovation Emerges from an Observation

An observation does not contain an unambiguous definition of the problem.

If projects repeatedly deliver late, this may be framed as a planning deficit. Equally plausible explanations include unclear decision-making rights, conflicting objectives, poorly translated requirements, unrealistic portfolio planning or dependencies that do not appear in the official process model.

Every frame directs attention towards particular causes and makes certain solutions more likely. Anyone who frames a structural interface problem as an individual communication problem will offer training, even though decision-making routes or responsibilities may need to be changed.

Problem framing is therefore not a neutral description. It helps determine whose knowledge appears relevant, who is assigned responsibility and which interventions are regarded as legitimate. Studies of complex public and social innovation initiatives show that experienced design teams continuously broaden and challenge problem frames and align them with the emerging solution space.

The Controversy: Does Innovation Always Begin with a Problem?

The demand to begin with a customer problem is sensible in many contexts. It protects organisations from developing solutions primarily because they are technically possible.

It is not, however, a universally valid rule of innovation. Innovation may equally begin with a scientific discovery, a new technology, a regulatory change, a previously unused resource or an unexpected combination of existing capabilities.

Even in these cases, an organisation must eventually clarify which relevant opportunity, tension or change requirement is involved. Innovation does not therefore have to begin with an already recognised problem. It does, however, require a sufficiently robust frame within which a new approach can be assessed and further developed.

3. Activating, Combining and Incubating Knowledge

Ideas do not emerge from nothing. They draw on existing expertise, experience, analogies and observations and recombine them in new ways.

The required knowledge is, however, distributed throughout organisations. Technical expertise exists in a different place from user experience. Senior managers understand strategic dependencies but do not necessarily know the informal steps through which employees keep existing systems operational. Customers may be able to describe difficulties without knowing their cause.

An organisation’s capacity for innovation therefore depends not only on how much knowledge it possesses, but on whether that knowledge can be activated and connected for the question at hand.

Incubation also matters. A meta-analysis of studies into creative problem-solving found a positive average effect from interruptions. The effect varied, however, depending on factors including the type of task, the preceding engagement with it and the design of the interruption. Incubation is therefore not a reliable magic trick, but it does provide empirically supported evidence that immediate time pressure does not improve every kind of thinking task.

The relevant question is: Where in our organisation can knowledge be connected and processed before a decision is already expected?

4. Idea Generation: Expanding the Space of Possibilities Without Confusing Quantity with Quality

Idea generation initially serves to expand the space of possible answers. Analogies, deliberate changes of perspective, altered constraints and unusual combinations can all be useful.

A large number of ideas is not, however, an end in itself. The assumption that greater quantity automatically produces better solutions oversimplifies the relationship. More alternatives may increase the chance of finding a strong approach. They also increase the evaluation burden and offer little value if every alternative is based on the same inadequate problem frame.

Group brainstorming is not automatically superior to individual idea development either. Groups can create production blocking: while one person is speaking, others cannot express their thoughts or adapt their contributions to what has already been said. A more effective structure may therefore combine individual preparation with subsequent collective development.

Research into workplace creativity broadly emphasises the interaction between the individual, the task and the social context. Creativity is neither exclusively a stable personal characteristic nor the reliable outcome of any single method.

A Current AI Controversy

Generative language models, or large language models (LLMs), can make ideas more accessible. In an experiment by Doshi and Hauser, participants wrote short stories either without support from a LLM or with the help of starting ideas generated by one. Access to the models improved the average ratings of the stories, particularly for participants whose initial performance was lower. At the same time, stories produced with model support became more similar to one another.

This does not demonstrate that AI (artificial intelligence) generally improves organisational innovation or inevitably reduces collective diversity. The study examined creative writing under controlled conditions, not the development of complex products, processes or business models. It nevertheless points to a relevant trade-off: a tool can improve individual performance while simultaneously creating shared starting points that narrow the collective solution space. AI can increase the number of available answers without broadening the range of questions being asked.

Organisations should therefore assess AI not only by whether people produce ideas more quickly. It is equally important to determine whether independent observations and unusual bodies of knowledge enter the process before a model begins offering plausible standard answers.

5. Elaborating Different Perspectives

Different ways of thinking initially broaden only the potentially available pool of information. They do not automatically produce better decisions or innovation.

Research into team diversity describes two opposing mechanisms. Differences can make task-relevant knowledge and varied perspectives available. At the same time, they may contribute to social categorisation, misunderstandings, conflict or reduced cohesion. Meta-analyses and reviews therefore show no simple, context-independent relationship between diversity and team performance. The type of task, the nature of the differences and the social processes involved all moderate the outcomes.

The decisive point is that information and perspectives must be exchanged, explained, examined and related to one another. Neither functional diversity nor access to external knowledge guarantees that this process will occur. One study of 215 knowledge-intensive teams, for example, found support for a relationship between external information search and elaboration, but not for a corresponding direct effect of functional diversity.

A recent meta-analysis of 38,304 teams also found an association between demographic diversity and stronger categorisation processes, but no general relationship with information elaboration. This does not mean diversity is without value. It means that differences do not automatically become knowledge that can be used collectively.

The central organisational question is: Are dissenting perspectives merely heard, or are they shared in sufficient depth for their internal logic to be understood, examined and connected with other knowledge?

6. How to Evaluate Ideas: Recognising, Comparing and Selecting Them

A good idea must do more than emerge. It must also be recognised as good. This is more difficult than many selection processes suggest. Originality, expected benefit, feasibility, strategic fit, risk and scalability are different criteria. An approach may be highly novel but currently difficult to implement. An easily implemented idea may change so little that its innovative value remains limited.

Research reviews on idea recognition and selection conclude that there is considerable room for improvement, particularly in the selection of creative ideas. Novel approaches create uncertainty, fit less readily into existing categories and are therefore easily rejected in favour of more familiar alternatives.

Nor is the idea alone being evaluated. The status, specialist language, presentation style and organisational position of the person proposing it influence how much attention and credibility it receives. An idea may be substantively compatible with an organisation and still lack social compatibility.

The implication is not that organisations should attempt to make evaluation completely objective. Under conditions of uncertainty, this is hardly possible. It is more useful to make criteria, assumptions and areas of ignorance visible. Rather than selecting a single winning idea too early, organisations can deliberately investigate several approaches with different risk profiles.

An uncomfortable diagnostic question is: Do we evaluate the potential value of an idea, or mainly its familiarity and the credibility of the person presenting it?

7. Generating New Insights Through Prototypes

A prototype is not simply an incomplete version of the eventual solution. It is a question given material form.

A clickable model can test whether users understand a particular logic. A role play can test decision-making routes. A manually delivered service can reveal whether a relevant user need exists at all. A technical demonstrator answers different questions from a pilot conducted under real organisational conditions.

The value of a prototype therefore depends on which uncertainty it is intended to reduce. If this is not clarified, impressive demonstrations may be produced without generating reliable grounds for decisions. Even a successful test proves only what was examined under the relevant conditions. It does not automatically demonstrate technical scalability, economic viability, acceptance, regulatory robustness or integration into existing processes.

Experiments also generate organisational learning only when assumptions are stated in advance, outcomes are subsequently evaluated and consequences are drawn. A failed attempt is not automatically valuable. Without evaluation, it remains merely a failed attempt.

Design Thinking in the Innovation Process: Useful, but Not a Magic Formula

Design Thinking combines several useful practices: close engagement with contexts of use, problem framing, interdisciplinary collaboration, visualisation, idea development and prototyping.

A systematic literature review by Rösch, Tiberius and Kraus examined 164 academic publications. It showed that Design Thinking is conceptualised in widely differing ways and that its application depends on contextual conditions, process design and organisational integration.

This explains why the question ‘Does Design Thinking work?’ is too broad. Design Thinking can provide a useful framework for particular forms of exploration and learning. A workshop does not, however, automatically answer:

• who will provide the resources required for implementation;
• which idea should take priority within the portfolio;
• how technical and regulatory dependencies will be addressed;
• which power interests may oppose the change;
• or how a solution will become permanently embedded in routines.

Design Thinking can strengthen parts of the innovation process. It does not replace organisational innovation capability.

8. Implementation, Use and Institutional Learning

This is where it is decided whether a potential new approach actually becomes an innovation. A solution must be integrated into technical systems, workflows, responsibilities, budgets and decision-making routes. People must not merely know about it; they must be able and willing to use it in the relevant context. Existing incentives must not make the old way of working more attractive than the new one.

A systematic review of innovation implementation groups the relevant influences into four areas: characteristics of the innovation itself, social factors, organisational conditions and individual factors. It also points out that innovations may change during implementation and that implementation is not merely the mechanical transfer of a finished concept.

A further research review identifies factors including a positive implementation climate, management support, available financial resources and a learning orientation as conditions that increase the likelihood of successful implementation.

The recursive process now closes its loop. Use produces new observations. Local adaptations may reveal that the original design was incomplete. Unexpected side effects require the problem to be reframed. An innovation does not merely change an object or a process; it can also generate new knowledge about how the organisation actually operates.

The decisive question is: What must change in routines, decisions, systems and responsibilities for the new approach not merely to be introduced, but to be used effectively?

Where Does Your Organisation Lose Insight?

Innovation capability is not demonstrated solely by research budgets, creativity workshops or the number of submitted ideas. It is demonstrated by whether an organisation can carry insights through all eight fields and continue developing them in the process.

The question is therefore not only whether a defined innovation process exists, but also where insight is lost. Are relevant deviations perceived? Are alternative problem frames permitted? Can distributed knowledge be connected? Do genuinely different possibilities emerge? Are perspectives elaborated or merely collected? Can unusual ideas be recognised fairly? Do prototypes answer questions that are relevant to decisions? And does the new approach ultimately change actual use?

Innovation often begins earlier than organisations assume and ends later than many innovation methods suggest.

An idea is not yet an innovation. It is initially only a possible answer to a problem that may itself have been misunderstood. Genuine innovation emerges only when an organisation can translate signals into insight, insight into testable possibilities and viable possibilities into actual use.

---

Under what social and organisational conditions lead different ways of thinking to innovation? Organisations already have access to a great diversity of perspectives and ways of thinking. Currently, most processes don’t make use of this. Monika Wolff explores and shapes the conditions needed to turn this diversity into shared insight, better decisions and genuine innovation. A two-week assessment and a report with recommended actions will give you a overview of what is possible (book dicsovery call).