Ανάρτηση Ερευνητικού Δοκιμίου no 06/23

Ερευνητικό Δοκίμιο no 06/23 με τίτλο "What Makes for Survival? Key Characteristics of Greek Incubated Early-Stage Startup(per)s during the Crisis: a Multivariate and Machine Learning Approach" των Ιωάννη Μπέση, Ιωάννας Σαπφoύς Πεπελάση και Σπύρου Παρασκευά

Περίληψη

This paper maps and analyses the total population of incubated startup(per)s in Athens, the main center of the startup ecosystem in Greece during 2010-2016- the worse years of the economic crisis. Its purpose is to uncover data for the key features of this enterprise group and the factors behind success as measured via survival. We use a descriptive statistics methodology for the total populations of 300 startups and in specific regarding the survival subcohort we applied a combination of a feature selection algorithm and tested a logistic regression modeling family. The basic findings of the descriptive statistics analysis for the survival subcohort (in comparison to the total population of our data base) show that during the crisis years 2010-2016:
Regarding founder specific indicators, a larger share of founders were: More mature; male; more educated; had degrees in economics and business, science and theoretical studies; were holders of degrees related to the startup sector and had more experience abroad.
Regarding startup/firm specific indicators: The share of Athens was lower and the share of the UK and the USA was higher. The share of Construction-Engineering and Transportation was higher. The share of srvices was lower. Notably, high tech goods and processes were higher and B2C transaction type was lower. The share of startups with customers abroad was higher and the share of startups with family members was roughly the same.
In our logistic regression model using MRMR algorithm, from the five most important input variables, the following had a positive impact for a startup‟s survival, i.e. a) providing services to both Customers and Businesses, b) having achieved customers abroad, c) having founding members that digressed from original studies and d) applying high-tech processes internally. Also, there was one variable identified affecting negatively survival, i.e the variance in Educational Level amongst founding members.

O  Ιωάννης Μπέσης είναι Business Development Manager στην a-Quant, η Ιωάννα Σαπφώ Πεπελάση είναι Ομότιμη Καθηγήτρια του τμήματος Οικονομικής Επιστήμης του Οικονομικού Πανεπιστημίου Αθηνών και ο Σπύρος Παρασκευάς είναι Machine Learning Engineer στην SPhears Al and Hellenic Air Force (HAF).