The presentation given at the 3rd international conference "Entrepreneurship, innovation and regional development", September 24-25, 2015. ISM University of Management and Economics, Vilnius
Linas Eriksonas, Social networks of startup entrepreneurs: the case of the startup ecosystem in Lithuania
1. Social networks of startupentrepreneurs:
the case of the startup ecosystem in Lithuania
Dr Linas Eriksonas
International Business School at Vilnius University
Linas.Eriksonas@tvm.vu.lt
2. Content
• Research background and aims
• Hypotheses
• Setup of the experimental study
• Results
• Implications for path-dependence theory
3. Background and aims
• Forbes (June 2015): Vilnius - one of five entrepreneurial hotspots in
Europe (alongside Eindhoven, Budapest, Tallinn and Lisbon):
• “a fertile innovation hotspot for trailblazing entrepreneurs”
• “the proof lies in its burgeoning startup communityin which over 100 million
USD has been invested in recent years”
• Aims of the research (initiated in 2012):
• to explain the factors which have been driving the growth of startup networks
in Lithuania
• to look for conditions that could create a path-dependence for sustaining the
momentum
• to design a research-driven model for a longitudinal tracking of such networks
4. What is a startup?
• A collective of people characterisedby informalarrangments of work,
shared lifestyles,hobbies, values and the ambition to become rich
• you can think of a startup as a way to compress your whole working life into a few
years. Instead of working at a low intensity for forty years, you work as hard as you
possibly can for four. This pays especially well in technology, where you earn a
premium for working fast (Paul Graham, Hackers and Painters, 2004)
• Venture labour or entrepreneurial workers (Neff,2012)
• Graduates, especially from business schools, without adequate possibilities on the
job market (first emerged in 2000-2001, then after 2008-2009) or after a career spell
in the corporate world
• Skilled in management and STEM disciplines, technologically adept
• Well integrated socially through alumni networks
• Their value system informed by different strands of thought originating in subculture
(hackerism, liberatarianism, humanism) characterised by the dream-like ambition to
succeed and change the world while taking execesive risks
5. Hypotheses
• Hypothesis 1: the growth of the startup networks in Lithuania has
been driven by individuals having initial social capital but being
without adequate employment opportunities on the job market
following the financial crisis in 2008-2009.
• Hypothesis 2: the startup networks were largely shaped by the public
and private funds which became available for networking activities
and raising initial funds for pre-seed/seed capital.
• Hypothesis 3: the growth of entrepreneurial networks have been
impacted by the use of social media, mainly Twitter which is one of
the main platforms for global communication among entrepreneurs.
6. Research design
• The main part of the research was carried during three periods in 3 years
(June 2013,October 2014 and September 2015) and consisted of 5 steps:
1. 20 experts have been identified and surveyed (using a respondent-driven
snowballing sampling method) which helped to establish the initial seed list of 51
startups and 28 facilitators for analysis;
2. A consolidated list of startups and facilitators nominated each by more than 25% of
the experts was drawn up and their social media accounts were collected
3. The data from the active Twitter accounts of 21 shortlisted subjects were collected
(using Python scripts): gathered data about over 20000 users having more than
22000 links; the core network analysed included some 1200 individuals.
4. The Twitter data were analysed and visualized with the help of software packages
for social network analysis.
5. In 2015 the analysis was updated which allowed to create the dataset of
longitudinal data and compare it with the initial dataset from 2013.
7. H1: growth of startup networks driven by
demand/supply on the job market
• The biographical analysisbased on Linkedindata and additional
input from publiclyavailablesources showed that >50% of
startup founders were graduateswithout steady jobs.
• The first generationof hackers (2002-2004) was absorbed by the
job market prior to 2008 (joined first IT companies which opened
business in Lithuania - Unity Technologies, Bentley Systems)
• The majority of the early startup founders
started their activitiesafter 2006 havingfew
years behindafter graduation(ca. 26 yrs)
• A typical founder: BA level educationin
management or IT, almost exclusively male,
with varied ethnic backgrounds
8. H2: the startup networks were largely shaped by the
public and private funds and their facilitators
Core Facilitators
Data from October 2014
Core Founders
9. Facilitators at the centre of the network
StartupHighway
acceleration
programme
Hub Vilnius
coworking centre
First startups
(2011-2012)
The acceleration
programme
StartupHighway
(at the co-working
hub at Northtown
technology park)
setup in the
emulationof
Seedcamp London
The first co-working centre in
Vilnius(HubVilnius) was setup
in 2011 in emulationof Riga‘s
TechHub, a franchise of
TechHub London
Data from June 2013
11. Results
• The startup networks have emerged after 2011 and have coincided with
the economic crisis in the country resulting in high levels of youth
unemployment among skilled or semi-skilled graduates and a number of
jobless junior-management level professionals)
• The major role in setting up the networks was played by publicly and
privately supported facilitators and their institutions: the co-working place
and the accelleration programme.
• The patterns were emulated by transfering practices from London
(Seedcamp, Open Coffee Club), via Riga (Riga TechHub) and Tallinn
(Garage48), early links with Helsinki (Arctic Startup).
• Update: (since 2014) direct links to RocketSpace (San Francisco) under
development, the transferred practices back to London (Drinkpreneur Live)
12. Implications for application of path-dependence theory
• The concept of imprinting (Marquis, Tilcsik, 2013):
• nascent research reflects the notion that “a past network, with its
accumulated relational experience, becomes a kind of ‘network memory’
• Despite this recent development, much remains to be learned about network
imprinting and, more generally, the lingering effects of network history
• Two factors can be considered using the longitudinal study:
• Importance of past ties for new knowledge, practices and routines (creating
career advantages)
• Structure of a network sustains the imprint of conditions (new network
entrants are likely to imitate the existing structural patterns)
13. • Average Degree: 1,414
• Modularity: 0,724
• Modularity with resolution: 0,724
• Number of Communities: 11?
• Diameter: 6
• Average Clustering Coefficient*: 0,070
• Average Path length#: 3.336760991859381
• Number of shortest paths: 45367802
Sept.
2015
June
2013
• Average Degree: 1,166
• Modularity: 0,741
• Modularity with resolution: 0,741
• Number of Communities: 12
• Diameter: 8
• Average Clustering Coefficient: 0,019
• Average Path length: 4.180968027664934
• Number of shortest paths: 64889943
14. Imprinting via shared work practices
Hackerspaces
shared
workshop
tools and
work space
Startup incubators:
Team work in a
shared space
Coworking
spaces:
Individual work
places in a
shared work
space
DIY/
Makers
Self-employed
professionals
Hackers
Startup founders
15. Does path-dependence exist in startup networks?
• Hacker collective “demoscena”
in Vilnius, Kaunas (2002)
A hacker collective at GamejamLT 2002: NeARAZ, Oasis,
rtfb, Voblia, ReJ Teaman, ProNinja, OneHalf, simple
Viktoras Jucikas (hacker nickname BigtoP), Rytis
Vitkauskas, founders of Yplan (founded in 2012)
• One of the ex-hackers pitching
startup Yplan in London (2014)
Funding received for Yplan:
$37.7M in 3 Rounds from 16 Investors
16. Paid career job
Work-in-progress: path-dependence model for
analysing startup entrepreneurs longitudinally
• Involvement in
subcultures
Hackers,
DIY/Makers
• Professional
career track
• Entreprenerial
activities
Startup teamChanges in personal
communitynetworks
17. Acknowledgements
• The research was funded by the Research Council of Lithuania (2012-
2014) as part of the group research project on innovation networks
• The access to Twitter data via Twitter API granted for development
purposes by Twitter
• Software used:
• Python scripts for Twitter, courtesy of Dr Derek Ruths, Network Dynamics Lab,
School of Computer Science, McGill University
• Neo4J database, the academic license granted by Neo Technology Inc.
• Gephi, an open source software (licensed by the Gephi Consortium)
• Visone, a free software for academic purposes (University of Konstanz)
18. References
• Eriksonas, L. 2010. Peer Driven Survey Methodology and Indicators
for Policy Relevant Research Competence Assessment [poster
presentation]. European Network of Indicators Designers Indicators
Conference, Paris.
• Eriksonas, L. 2013. The Impact of Time Zone Difference on Social
Networks of Entrepreneurs. Sunbelt 2013 Conference, University of
Hamburg, Hamburg
• L. Eriksonas et al., Inovacinių tinklų kūrimasis Lietuvoje: atvejų studijos
(Innovation networks in Lithuania: case studies) (Vilnius, 2015)