Research in industry may take several roles, and each of the roles require different kind of organization and type of leaders.
Role one: Research as the "smart arm" of the enterprise: In this case, the research activities are part of the daily (tactical and operational) business of the organization. Research people are going to pre-sale meetings, doing PoCs for customers, providing services to customers, and help product organizations in getting new features to products, typically in cases where knowledge and skills that they possess are needed. The research organization in this case is geared towards the short-term impact on the organization, and is measured by its tangible impact. In this case the research leadership should be geared towards good knowledge of the enterprise business and also well connected internally, to trace opportunities for getting more traction for business; the research is being funded by the business units according to its contribution, and the research leaders should be able to smoothly be able to switch jobs with any other leader in the organization, as organizational business and networking is the main qualification.
Role two: Research as the "scientific arm" of the enterprise: In this case, the research activities are geared to advance science and to research areas that are geared towards impact on the knowledge in the universe, and on long term directions of the enterprise. One of the questions is what is the tangible benefit of the enterprise from this type of research has been investigated recently to view what is the $value on the brand image, as a function of the reputation of research done there, this is also reported in the IBM Zurich Lab Director Blog. In this case the evaluation can be according to research quality metrics common in the academic world, as well as the impact of the research on corporate. In IBM Research the relational database is often mentioned as the most influential research projects both on the science and on IBM business. Funding can come from the enterprise as well as from external and government oriented research grants, similar to academic research funding. In this case the research leaders should be able to switch jobs with Deans of relevant faculties in top universities, or head of independent research institutes, as good familiarity with the research community and ability to lead scientific work, and good acquaintance with the research grant business.
Role three: Research as the "incubation arm" of the enterprise; In this case, the research activities should be constructed as a collection of start-ups inside the enterprise. Each of them gets some seed money and gets to its way, having the ability to sell and gain customers, in some cases even get external funding as partners. As in any incubator, some of the start-ups may fail, some of them may exit, and develop spin-offs out of the enterprise, and some of them grow inside the enterprise and become business units, in a similar model to the model of acquisition of other start-ups; this is an interesting model that typically enterprises don't really know how to apply, but there were examples of technologies, companies, and business units within the same enterprises that were originated in such start-ups. This also distribute the leadership, since each such start-up requires a leader with intrapreneurship abilities. The funding model is of start-ups where the enterprise is acting as VC, and the overall leaders should be able to switch jobs with VC executives, and they also need to create an atmosphere that leverages the enterprise resources where it helps, and masks the enterprise when it interrupts; The leader of each individual start-up leaders should be able to switch jobs with external start-up leaders (though inside start-up also requires some internal political skills).
Current research organizations try to have a mix of all these, the main issue is the diversity of leadership skills and how work should be organized which is different in each of them vs. the tendency of organizations to have homogeneity both in the organization type and in type of leaders.