Following the identification of the project's scientific or technological advance and associated uncertainty, it is essential to quantify which costs are included in the R&D Tax Credits claim. The scope of a project for R&D purposes will most often be different to that of the Project for your business purposes.
"The activities which directly contribute to achieving this Advance in Science or Technology through the resolution of scientific or technological uncertainty are R&D."
The includable costs are those which are required to overcome Technological or Scientific Uncertainty directly (but also see QIAs). Thus, technical issues that cannot be addressed by off the shelf information alone. The R&D project starts when work to resolve the Uncertainty commences and ends when the Uncertainty is resolved or work otherwise ceases. Thus the end of the Project may be when you have a working prototype or when the technology goes into production, or if you decide not to take the project forward.
It may be that after the initial Scientific or Technological Uncertainties have ended (and possibly that the product has moved into a production environment) that new Uncertainties arise. Such new Uncertainties would mean that the R&D process may start again. However, you need to be clear about which problems are overcoming Scientific and Technological Uncertainties and which are for more routine maintenance of technology.
From a roles perspective, Developing, Testing, Managing the Development are examples of activities that directly contribute to overcoming the Technological Uncertainty.
From a components perspective, it may be that a team is developing a component that is routine in itself but directly contributing to a broader R&D project. In this case, the work to create the standard component is also R&D for Tax Purposes. For example, in Software Development, this could be the case if a company is developing a user interface, that is itself routine, but is required to import data into the back end of the platform, which contained the Technological Uncertainty. Hence without the User Interface, there would be no way to test the Uncertainty of back-end of the platform, meaning that the UI (at least the development of the barebones import capability) will directly qualify as R&D.