03 · What You Need to Know
Where Research Ends and Innovation or Development Begins
Research and innovation are related but not synonymous
Research can generate knowledge that eventually enables innovation. Innovation can create problems that require new research. Organizations may deliberately combine research, engineering, design, development, and implementation within the same project.
Still, research and innovation answer different primary questions.
Research asks, broadly, What can we learn? Innovation asks, What new or improved product or process can be put into use?
The OECD's Oslo Manual 2018 defines an innovation as a new or improved product or process, or combination of the two, that differs significantly from the unit's previous products or processes and has been made available to potential users or brought into use.
Implementation is therefore important to innovation under this framework. An idea that remains only an idea is not yet an innovation in this measurement sense.
Research can occur without innovation
A study may develop important knowledge without creating a new product or process. Researchers may investigate a theoretical question, document a natural phenomenon, replicate an earlier finding, analyze historical evidence, estimate a relationship, or develop an explanation.
The resulting knowledge may eventually support innovation, but it does not have to.
This follows directly from what makes an activity research. The defining concern is systematic knowledge development, not whether a usable artifact emerges at the end.
Innovation can occur without research
The reverse is equally important. An organization can introduce a new or significantly improved product or process by applying knowledge that already exists.
The Oslo Manual explicitly recognizes that R&D is only one of several activities through which knowledge useful for innovation may be generated or acquired. Organizations may also draw on engineering, market research, user data, design, training, software development, and other activities.
An organization therefore does not need to conduct original research every time it innovates.
Suppose a university combines well-established technologies into a new student-advising platform. The platform may be innovative for the institution even if its developers solve no unresolved research problem and generate no additional scholarly knowledge.
Invention and innovation are not identical either
An invention may involve the creation of a novel device, technique, or idea. Innovation requires more than novelty under the OECD framework: the new or improved product must be made available to potential users, or the process must be brought into use.
This distinction helps explain why a prototype can be inventive without yet becoming an innovation.
Research may contribute to invention, invention may contribute to innovation, and innovation may generate new research questions, but the concepts should not be collapsed into one another.
Development is a broad term
The word development causes particular confusion because it is used in many ways. Software development, curriculum development, product development, organizational development, and experimental development do not mean the same thing.
A developer can build a product using established knowledge. That activity may involve considerable design, engineering, testing, and iteration without necessarily qualifying as research.
Research and experimental development, or R&D, has a more specific meaning in research statistics.
Experimental development is part of R&D under the OECD framework
The OECD's Frascati Manual divides R&D into basic research, applied research, and experimental development.
Experimental development is systematic work drawing on knowledge gained from research and practical experience, producing additional knowledge, and directed toward producing new products or processes or improving existing ones.
The phrase producing additional knowledge is crucial. Experimental development is not simply the routine act of constructing the final product.
For an activity to qualify as R&D under the Frascati framework, it must satisfy five criteria: it must be novel, creative, address an uncertain outcome, be systematic, and be transferable and/or reproducible.
Experimental development should not be confused with product development
The Frascati Manual explicitly distinguishes experimental development from product development.
Product development is the broader process extending from ideas and concepts toward bringing a product to market. Experimental development may occur during that process when unresolved knowledge or technical uncertainty requires systematic work that generates additional knowledge.
Once the relevant R&D uncertainty has been resolved, later engineering, production preparation, routine software work, marketing, commercialization, or implementation may continue without themselves constituting R&D.
Research
Systematic investigation directed toward developing, testing, refining, reinterpreting, or applying knowledge.
Experimental development
Systematic R&D work that draws on existing knowledge, produces additional knowledge, and is directed toward new or improved products or processes.
Innovation
A new or significantly improved product or process that has been made available to potential users or brought into use.
Product development
The broader process of creating and bringing a product toward implementation or market, which may include R&D as well as many non-R&D activities.
Technical uncertainty is an important clue in R&D
Imagine two software projects.
In the first, developers build a standard registration system using technologies and architectures they already understand. The implementation is difficult and takes thousands of hours, but the central technical problems have established solutions.
In the second, a team encounters an unresolved technical problem for which existing knowledge does not provide an adequate solution. They formulate alternative approaches, develop prototypes, conduct systematic tests, analyze failures, and generate transferable knowledge about which architecture can solve the problem.
Both projects involve development. The second has a much stronger R&D character because it addresses genuine uncertainty through systematic knowledge-producing work.
Difficulty alone is therefore not the same as research uncertainty. A task can be extremely difficult because it is large, tedious, expensive, or technically demanding while still relying on known methods.
Building an artifact can be part of a research methodology
In engineering, computing, design-oriented disciplines, and educational technology, researchers may create artifacts as an integral part of research.
A prototype can embody a hypothesis about how a problem might be solved. Building and evaluating it can generate knowledge about design principles, mechanisms, technical feasibility, user interaction, or performance under specified conditions.
The artifact itself, however, is not the entire research contribution. Researchers should be able to explain what knowledge problem the artifact addresses, why the design decisions matter, how the artifact is evaluated, what evidence is generated, and what can be learned beyond the fact that the artifact was successfully built.
A new application is not automatically a research contribution
Suppose a student develops a mobile attendance application using a standard framework, database, authentication system, and QR-code library. The application may be useful and professionally executed.
If the project merely demonstrates that these established components can be assembled into a functioning system, its novelty as a local product does not necessarily establish a research contribution.
Research would require an additional knowledge question. Perhaps the project investigates a novel privacy-preserving attendance method, compares competing interaction designs under specified conditions, or develops and evaluates an algorithm addressing an unresolved technical limitation.
This is the same distinction encountered when asking what counts as new knowledge in research. Something can be new to the developer or institution without being a meaningful new contribution to knowledge.
Testing whether a product works is not automatically research either
Developers routinely test software, devices, processes, and services. They conduct unit tests, usability tests, quality assurance, performance checks, safety assessments, and acceptance testing.
Testing is necessary for development, but the word test does not automatically make the activity research.
If a team tests whether an application meets predetermined technical requirements, that may primarily be quality assurance. If researchers systematically investigate an unresolved question about how a novel design performs and use the results to produce transferable knowledge, the activity may have a research component.
Again, purpose and knowledge contribution matter more than the visible technique.
Implementation is central to innovation but not required for research
A research project can conclude with a theoretical model, empirical finding, experimental prototype, or unsuccessful attempt that nevertheless produces useful knowledge.
Innovation, under the OECD definition, requires implementation. A product innovation must be made available to potential users, while a process innovation must be brought into use.
This produces an important asymmetry: research can succeed intellectually even when no usable product emerges, whereas an unimplemented idea does not satisfy the Oslo Manual's operational definition of innovation.
An innovation does not have to be successful to count as innovation
Implementation does not guarantee benefit.
The Oslo Manual notes that an innovation need not be commercially, financially, or strategically successful. Nor does the definition require the innovation to produce a positive societal outcome.
This is useful because it separates classification from evaluation. Whether something is an innovation and whether it is a good innovation are different questions.
Research and innovation can form a cycle rather than a pipeline
The relationship is sometimes portrayed as a neat progression: basic research produces applied research, applied research produces development, development produces innovation, and innovation reaches users.
Real innovation systems are less obedient.
Practical problems can stimulate fundamental research. Prototype failures can expose theoretical gaps. User behavior can generate research questions. Research tools developed for one purpose can become innovations elsewhere. Implementation can reveal problems that send teams back into experimental development.
The OECD innovation framework likewise treats R&D as one among multiple knowledge-based activities that can contribute to innovation rather than as a mandatory first stage of every innovation.
Research status and commercial value are separate questions
A research project does not become more legitimate because it produces a profitable technology. Nor does commercially valuable development automatically become research.
Universities and firms may value research, intellectual property, prototypes, products, commercialization, and societal impact for different reasons. Those outcomes can overlap, but they should not be treated as interchangeable indicators of research quality.
The research question remains: what defensible knowledge did the investigation produce?
