How Concentrated Are Innovation Networks?

September 14, 2017

The 2017 R&D Trends Forecast by the Industrial Research Institute estimates that global R&D investment will increase by 3.4% in 2017 to $2.066 trillion with academic institutions and federal labs making up a significant portion of that total. Government and academic researchers play a critical role in developing new technologies, and one would expect that companies seeking to identify and acquire external technologies would maintain highly diversified networks of partners that span scientific specialties, university systems, and geographic regions.

Diversity in innovation networks depends on maintaining active research partnerships and collaborations with universities and organizations outside of their geographic region and national boundary. While scientific research tends to follow similar trends globally, there can be wildly different levels of intensity and specialization in technological areas, meaning that organizations would be largely unable to depend on a handful of partnerships to sustain an innovation network.

Although many companies believe they maintain a very diversified innovation network, this is not always true.





Wellspring Scout uses machine learning to scan hundreds of public and proprietary data sources to stitch together profiles on innovation partnerships. Using Scout, we analyzed the technology partnership data for ten leading companies in a variety of industries.

Specifically, we wanted to see whether external research collaborations were concentrated among a handful of partners or if organizations were making efforts to diffuse their partnerships across a number of diverse sources, geographies, and fields of research. In order to quantify research collaborations, we recorded the number of joint research publications and joint patents that these companies produced with external universities or government institutions.

As a whole, we found a great degree of concentration within innovation networks of these large companies. For our first example, BASF has hundreds of technology scouts, thousands of employees, and likely dozens of ongoing innovation projects at any given time. BASF, according to Wellspring intelligent search data, partners with the same five organizations on 15% of all jointly published research. Furthermore, they also work with the same five organizations on 63% of all jointly filed patents. Expanding the scope to the top ten collaboration partners, we see even greater concentration, with the top ten partners accounting for 24% of publications and 72% of all patents.

The trend holds largely the same across the other large companies we sampled. In other words, the largest and most innovative companies in the world complete a third to a half of all technology collaborations with the same five partners. At FoxConn, over 96% of innovation collaborations occurs with the same ten partners. This begs the question, are these companies missing opportunities? Is innovation myopia preventing companies from fully exploiting the $2.4 trillion dollars in research conducted around the world every year?


Wellspring Scout data shows that many large organizations have a heavy bias towards a handful collaboration partners. On average, a corporation will concentrate over 60% of collaborations with just ten organizations. There are thousands of dedicated research organizations conducting ground breaking work all over the world, yet most companies only work with the partners they already know. 

Why the concentration?

While this is by no means a rigorous academic work, and our sample size was admittedly small, the data here suggests that organizational partnerships are formed based on relationships more so than on potential technology opportunities. Much of the collaboration activity was also concentrated within globally recognized universities (i.e. the University of California system, etc), with smaller institutions receiving fewer partnerships. While that may be due to a number of factors, I suspect that prestige plays a significant role in technology networks as well.

In the long-term, this can leave organizations with a myopic understanding of a technology area and vulnerable to sudden technological change. Although organizations take an educated risk that their current partners are on the forefront of their fields, this cannot always be relied upon.  Companies leave themselves vulnerable to disruption by depending on a handful of partnerships, as the diffused nature of university research and corporate R&D means that even the top schools and organizations do not have a monopoly on potentially disruptive research.

Even prestigious universities are not the top in every field of research, so universities that partner with elite institutions must be aware to focus on their actual technology needs rather than the public relations appeal of such a partnership. There is a significant regional bias that can be exhibited when examining these partnerships. Many organizations prefer working with universities and organizations within their geographic proximity due to ease of communication, legal incentives (regional governments can incentivize partnerships with local universities), and existing relationships (large enterprises and local universities often have symbiotic relationships).

Some of this is a matter of geographic proximity: relationships are easier to maintain when they are geographically close, meaning companies will naturally prefer partnering with institutions that are within the shortest distance possible. This isn’t a unique problem in knowledge networks, as this effect can be seen in areas such as venture capital funding -- investors will generally prefer to meet with companies that are within a close drive, which is why so much venture capital funding is centered in major metropolitan areas.

Organizations are leaving opportunities on the table by depending on their established relationships to furnish technologies. Given that research funding is widely spread out between universities in multiple countries and regions, there is no good reason for organizations not to be establishing and maintaining relationships across the entire spectrum of research institutions.

While difficult, these challenges are not insurmountable and can be properly addressed if enough focus is given by an organization. Companies can diversify their knowledge supply chain by creating more geographically diverse networks through research partnerships, licensing opportunities, and professional events such as conferences. Using search tools, organizations can diversify their scouting efforts by identifying new partners rather than depending on old relationships to maintain their innovation networks.




Let's Talk