Innovation does not necessarily mean R&D

A business may launch a product that is new to its customers, digitalise a process or install advanced technology and clearly be innovating. Those activities may nevertheless not constitute research and development.

The distinction matters when managing projects, recording expenditure and preparing SIFIDE or grant applications. Treating all technological investment as R&D creates credibility risk. Excluding genuine experimental work may cause the business to lose knowledge and potential support.

What characterises innovation

Innovation broadly concerns the introduction or implementation of new or significantly improved products or processes. It may result from adopting existing technology, organisational change, design, purchased knowledge or the business's own R&D.

A solution can be innovative for the business or its market without creating new scientific or technological knowledge. Implementing an ERP, automating a line with commercially available equipment or adapting a product to a customer segment may create value and transformation without automatically being R&D.

The five R&D criteria

The OECD Frascati Manual identifies five criteria that should be met by R&D activity:

  • novel, because it seeks new knowledge or a new application that is not an obvious solution;
  • creative, because it is based on original concepts, hypotheses or approaches;
  • uncertain, because the outcome, cost or technical route cannot be determined in advance with confidence;
  • systematic, because it follows a plan with resources, records and a method;
  • transferable or reproducible, because knowledge and results can be documented and used or reproduced.

No criterion should be considered in isolation. Organised work is not R&D if it merely applies a known solution. An accidental new outcome also does not demonstrate systematic research.

New to whom?

The fact that a solution is new to the business is insufficient where competent professionals could implement it using normally available knowledge. The assessment should consider the relevant state of the art in the sector and technological field.

Conversely, a project need not produce a world-first scientific discovery. Experimental development can combine existing knowledge in a non-obvious way to overcome specific technical limitations.

Uncertainty is often the clearest indicator

Asking “what do we not yet know how to do?” helps distinguish R&D from routine execution. The uncertainty must be technical rather than merely commercial, financial or timetable-related.

Not knowing whether customers will buy a product is market risk. Not knowing whether a material can withstand a specific extreme condition, despite existing testing and knowledge, may be technological uncertainty. Uncertainty about implementation time caused by limited resources does not turn the work into research.

Activities that are generally not R&D

Depending on context, the following will normally fall outside R&D:

  • acquiring and installing equipment available in the market;
  • standard configuration of business software;
  • aesthetic changes or routine adaptations;
  • ordinary bug fixing and maintenance;
  • market research and commercial campaigns;
  • routine production and quality control;
  • training without an experimental component.

These activities may be essential and innovative. The distinction does not diminish their value; it recognises that they serve different purposes.

An industrial example

Consider a business purchasing a faster machine to increase capacity. The investment may constitute productive innovation, but it does not necessarily involve R&D.

The position changes if the business develops a new process because available equipment and methods cannot work a particular material with the required precision and energy use. Hypotheses, prototypes, tests, failures and iterations may constitute experimental development where the criteria are evidenced.

A software and artificial-intelligence example

Integrating an artificial-intelligence service through an available API may be technology implementation. Developing a method to achieve reliable performance where data is limited, biased or technically difficult and no known solution exists may involve R&D.

Use of advanced technology is not the criterion. What matters is an unresolved problem, uncertainty, creative work and a documented experimental method.

Separating activities within the same project

A business project may contain R&D, engineering, design, certification, industrialisation and commercialisation. It is not necessary to classify everything in the same way.

A sound structure identifies boundaries, owners, timetable and costs for each component. This improves management, supports selection of the appropriate funding instrument and reduces the risk of including expenditure without a direct connection to research.

Documenting the assessment

The business should be able to explain the state of the art, uncertainty, hypotheses, experiments, results and knowledge gained. Technical records, versions, test reports, minutes, timesheets and decisions help demonstrate that the work was systematic.

At Fenix Capital Partners, we begin with the technical substance and the reality of the project. Only then do we assess potential eligibility for SIFIDE, Portugal 2030 or ANI recognition. This sequence protects application quality and consistency between narrative, resources and expenditure.