Issues to consider when managing innovation: example of Intel’s lablets
I am still a bit obsessed on how companies manage their innovation processes and how they make it fit with their culture. In this context, I recently read an HBS case about Intel research. The “Intel lablets” particularly attracted my attention and raised general questions about the management of innovation.
In 2004, Intel realized that faster and more powerful processors will not be the most important customers’ need anymore and that it had to shoulder more of the burden of innovation. But Intel’s historical approach to R&D didn’t allow the company to spot the technologies that can impact the core business dramatically, or create new businesses. Indeed, Intel has based his R&D strategy on the idea of connecting closely research with manufacturing, by allocating R&D resources to each business units. This structure, even if really efficient in sustaining Moore’s law, was only focused on incremental innovations in the silicon roadmap, making it inefficient to identify some disruptive technologies.
These technologies could only be developed through “exploratory” research. However, Intel’s aversion for a centralized lab, which is definitely the standard model for this kind of research, pushed the company to look for another model. The search for a “happy medium” led to the creation of the “Lablets”, which are some small labs located close to universities well known for their collaboration with the industry and for their expertise in some sectors of interest for Intel. They selected Berkeley, Pittsburg and Seattle. Some characteristics of these lablets are differentiating them from some firms’ previous attempts to build labs close to universities and are particularly conceived to foster innovation:
- Thanks to an Open Collaborative Agreement, much of the research can be published and shared widely, with Intel hoping to acquire proprietary advantage in the downstream stages of the projects;
- Lablets are staffed with 40 researchers and are co-directed by an Intel research employee and a university faculty member changing every two years;
- Projects in lablets are funded on a milestones basis and were reviewed quarterly.
But this clever approach raises a certain number of issues.
How can the scope of activities of an “exploratory innovation program” be defined?
If the lablets’ role is to sense the environment to understand potentially disruptive technologies, then the scope is really wide. So it has to be reduced to some large sectors likely to have an impact on Intel businesses or ecosystem, as the final objective is to end up by passing projects to business units. The model thus represents a tricky balancing act: giving researchers enough freedom to be creative and to think broadly in terms of sectors in which to spot disruptive technologies, while confining their efforts to predefined aspects. There is a happy medium to find between the systematic search for opportunities congruent with core capabilities and “adhocracy”. Intel’s answer to this problem is to have Intel specify the problems it wants researchers to solve, but not the idea it wants them to work on, which sounds like a really fine distinction.
What type of structure is the most appropriate for this kind of exploratory research?
When I heard about the lablets, I couldn’t help thinking: is it really the right structure to be in charge of the strategic responsibility of understanding what will be disruptive in the market? Having only 40 researchers in each of the 3 lablets is a good way to limit costs and therefore limit the impact of the high level of risk involved, which is what companies try to do most of the time with their innovation team, but it also can appear difficult to reach the critical mass, often mentioned as crucial to create a fertile environment. Moreover, even if being a small group can limit the internal resistance (as this kind of exploratory research is absolutely not part of Intel’s culture) by not attracting too much attention, it doesn’t allow stand on your own in a large organization. Intel lablets would require a heavyweight team (see matrix) to push projects into business units without having to fight to convince the business units of the viability of a project. In addition, the frequent turnover of people due to the university involvement can be a threat in the case of long term projects. Besides, if some lablets’ projects make it downstream into the main R&D fold, then some great researchers will probably leave to ensure the follow-up. But it seems that this is the price to pay to always bring new blood into an innovation program, which is crucial to generate more ideas and more expertise.
How to assess the performance of such structures?
What I find interesting here is that the lablets’ case reflects the basic problem of any innovation process: how can you set up performance metrics in a domain that flourishes only when there is a tolerance to failure? Finding metrics to measure the success of exploratory research is difficult as the process of sensing the environment can lead to many useless results, but is still necessary to have chance of producing breakthrough products or predicting disruptive changes in the market. It is clearly the case of R&D as a Real Options: you pay a certain amount of money to have the right to exercise an option when the market changes. Which is why funding the projects on a milestones basis is so important in this context, as this sequential investment creates the option to abandon the project in midstream if it is too early to market or doesn’t fit the company’s strategy. Each stage can be viewed as an option on the value of subsequent stages, and valued as a compound option.
The point of this post was not to speak about Intel’s initiative in particular but more to discuss about the issues that any innovation process will generate. So just for information, since their creation Intel’s lablets have been working more and more closely with business units to facilitate the downstream transmission of projects, and have generated some cross-industry collaborative projects, like PlanetLab. If you have been part of an innovation program or have any thoughts about these basic issues, do not hesitate to use the comments part!
Related posts:
Books, Business Process Management, Business intelligence, Business strategy, Consumer electronics, Entrepreneurship, Finance, Globalization, IBM, Internet, Organization, Project Management, Silicon Valley, Technology, innovation, operations
