Lectio Praecursoria: Analyzing Innovation Ecosystems As Networks

Jukka Huhtamäki
Tampere University of Technology
November 18, 2016

Analyzing innovation ecosystems is difficult.

Main innovation actors are individual persons, companies, advisers, and investors. Innovation activities have several stakeholders, including researchers studying innovation, policy makers, and universities as sources of knowledge supporting innovation.

Each of the aforementioned actors access multiple of sources of information. Media covers individual stories, such as the creation of Jaiku, an early Finnish social media startup that was sold to Google. Both founders, Jyri Engeström and Petteri Koponen continue to operate in the Finnish startup scene. Individual events are important.

Statistical data on the development of national-level innovation activities exists, of course. Key issues of the existing statistics are their general nature and long delay. An innovation policy maker is not able to easily gain actionable insights on basis of knowing the overall change of the amount of growth companies in Finland. In particular in case that information arrives with a delay in the range of 12 months.

Currently, Finland is phasing a dramatic drop in research and education funding. Both university and innovation policy actors have an urgent need to be able to analyze in detail the impact that this kind of a shock will have for the innovation activities. Most importantly, they need to be able to share their findings with others in a way that drives the message home.

At Tampere University of Technology, we are most proud of our company companies. A total of 180 companies employ more than 6000 people and have created more than 1.2 billion (1200 million) in revenue. One of the recent examples is Vincit. Vincit CEO Mikko Kuitunen, Tampere University of Technology alumnus, has received a several awards on his work. Recently, Vincit initial public offering gave Kuitunen and other founders means to move from entrepreneurs to investors, should the choose to do so.

Social media is a never endings source of data. Companies announce that they are hiring new individuals, prime ministers and presidents tweet about their visits to innovation incubators, and databases of startup activities are collected and curated in a crowdsourced manner. Although a lot of social media data is just noise, we are in the early phases of exploring ways to use it to identify signals.

Collecting and making sense of this type of fragmented information is a popular topic in fiction. From Homeland’s CIA agent Carrie Mathison to the movie character based on mathematician John Nash, the original developer of game theory, analysts seeking to find mechanisms underlying individual events and pieces of information have used walls of paper organized in an appropriate way. One particular item is found in most of these walls, that is, strips of woolen yarn connecting individual pieces of information, often based on the interaction between individuals. These walls are sometimes dubbed as “crazy walls”.

A systematic approach is needed to investigate innovation activities in ecosystem level. Innovation ecosystems are dynamic, interconnected, self-organizing ensembles where interdependent individual people, companies, investors, and other organizations come together to develop new products and services.

Innovation ecosystem analysis is particularly important to three group: innovation researchers, innovation policy makers, and the actors of the ecosystems, that is, companies, investors, and individual persons. Analysis is conducted in three levels, individual actors, pairs of relationships, and in the ecosystem level. Here, we focus on the ecosystem level.

Sensemaking is a key activity in the analysis. Several approaches exists to sensemaking, all of which stress the importance of an iterative and interactive approach. In sensemaking, individual events are observed and used to infer underlying mechanisms and structures. The network structure between the individuals is the most important individual structure.

Data on innovation ecosystems can be placed in three categories, data on actors, their affiliations (a person founds a company, advises a company, or works for a company), and transactions between actors (an investor invests into a company, a company acquires a company).

In network analysis, data on innovation ecosystem actors is used to construct a network representation of the ecosystem. Blue node is an individual that starts a company, a red node, another individual joins the company, and the company gets an investment. Similar events take place in the different parts of the ecosystem and appear in its network representation.

Before concluding the presentation, let’s review three innovation ecosystem examples, a local innovation ecosystem engager Demola, a European-wide EIT ICT Labs, and Silicon Valley. In the European case, we see that investors are the main connecting tissue between the individual cities that form EIT ICT Labs. Only a handful of individuals have worked in companies originating from more than one city. Interestingly, many of the venture capital investors are Silicon Valley-based. What if San Francisco Bay Area would be added into the network representation of this broad-based innovation ecosystem? Let’s try! Interestingly, it seems that SF Bay Area does locate in the middle of Europe. The densifications on top left are Google and Yahoo. The number of total nodes and edges is several times larger with San Francisco Bay Area included. We are also able to observe the evolution of an innovation ecosystem, here Demola. Nodes here represent project groups that are connected to the three universities on basis of student affiliations. Light green nodes represent companies that feed in their ideas to student based projects.

In order to manage the analysis process, a systemic approach is needed. We have designed the ostinato model to structure the process in a way that supports both the iterative and interactive approach imperative in the investigations and the automation of the process.

Dissertation is available online:

Huhtamäki, J. (2016). Ostinato Process Model for Visual Network Analytics: Experiments in Innovation Ecosystems. Tampere University of Technology. Vol. 1425. Available: http://urn.fi/URN:ISBN:978-952-15-3846-9