X-Ray Assessment of a Company R&D Organization
In order to assess the R&D program of a competitor, a potential partner or an acquisition candidate, it can be useful to understand the organizational structure of key R&D projects or programs. Under non-disclosure and in advanced stages of discussion, an organization may be willing to disclose the current or past organizational structure of a sensitive group such as a group within R&D, but often such information is very carefully guarded.
You can also gain organizational structure information from interviews, but you must find a sufficiently senior person who is willing to disclose information. The problem is that interviews of "talkative" employees come with ethical and legal risks too. However, a good approximation of the organizational structure of an R&D effort within an organization can be obtained using patent (or publication) data and co-author analysis.
To illustrate the method, Emerging Tech Insights conducted an x-ray analysis of the non-invasive glucose research efforts of a major healthcare company. Non-invasive glucose was chosen as the study subject because I once managed an R&D effort in the area and I can attest to the accuracy of the method in this case.
BACKGROUND For diabetics, maintaining their blood glucose within the narrow physiological range is difficult without multiple blood tests each day. Current technology uses a pin prick to draw a small sample of blood for each test. Six to ten pin pricks per day for decades of life is painful and inconvenient. For every diabetic, the holy-grail in blood glucose control is non-invasive testing. The idea of shining a light on a body part and somehow obtaining a blood glucose value would be more than valuable and would represent a true market disruption if it were available.
There is little doubt that non-invasive glucose, if available, would be a market success, but how can you assess a competitor's efforts in the area? You can certainly read their patents to determine their technological focus, but how can you assess their commitment to the area? You could a clear picture if you could understand the organizational structure of the research effort. Co-author analysis provides a way of making such an assessment without confidentiality or gray-areas into which interviews can stray.
DATABASE DEVELOPMENT To build a study database, we collected all patents for a specific healthcare company that mentioned the phrase "non-invasive glucose" anywhere in the patent. The database was ultimately composed of 135 unique authors and 269 patents. The study was completed in early July 2009.
CO-AUTHOR ANALYSIS In co-author analysis, one counts the number of times a pair of authors appears together in a collection of patents. Then, the results are visualized using a force-directed graph, and further analyzed by asking the computer to find clusters of authors.
GRAPH STRUCTURE In Fig. 1, the size of an author's circle depicts the number of patents authored. Authors that publish frequently together are placed near one another. Authors in green are those that I recall being a member of an internal venture. Authors in gray are authors I do not recall meeting despite searching for help within the organization. The red and blue circles represent leaders of the venture sub-groups.
OBSERVATIONS Examination of the Fig. 1 shows the existence of multiple functional groups, a finding corroborated by the cluster, Fig. 2.
From Fig. 2, it seems legitimate to conclude that the organization formed about five major groups and a few smaller groups to find and build a non-invasive glucose testing product. Notice (Fig 1. and Fig.2) that the two groups in the lower left corner seem to be coupled together through the light gray connecting lines. This hints at a combined group with shared resources. The cluster graph (Fig. 2) with its overlapping colors seems to confirm the observation. The cluster graph also hints that another fairly large group (composed of the red diamonds) was also formed to deal with some aspects of non-invasive glucose testing.
Just by counting the larger circles highlighted in Fig. 3 and by knowing the "cost" or "run-rate" for a scientist, one can estimate the budget of the group involved in one of the product development efforts.
30 people x $150,000 (overhead + salary + benefits) = $4.5 million (per year at the peak)
Analysis of the time frame of activity, a temporal picture, can further determine when a specific group was working and if there were competitive activities within the organization.
The patenting activity chart suggests that there were at least three major product development efforts in the organization. The first very small effort occurred (appeared in the patent literature) in 1992. A larger, much more sustained effort peaked between 2001 and 2003, and another large effort appeared later in about 2008. Since we know the author's names in the circled group from Fig. 3, we can determine when that group was active. Their work is highlighted in yellow in Fig. 4. Analysis of initial assignees shows that the last peak in Fig. 4 resulted from an acquisition.
CONCLUSION When vetting a partner, acquisition or competitor's interest in an area, co-author analysis can provide cost-effective, risk-free information about organizational structure as it was when publication occurred. Limited or focused interviews to confirm your understanding of an R&D program's structure are quite different and more effective than probing or unfocused market research questions.
None of the information we derived from co-author analysis is typically public information. However, with just a little analysis, we can see inside a competitor or acquisition candidate very clearly and very uniquely.