Analytics dashboards and user behavior: Evidence from GitHub

Schilpp, Jan ; Pethig, Florian ; Höhle, Hartmut

DOI:	https://doi.org/10.23919/MIPRO57284.2023.10159843
URL:	https://ieeexplore.ieee.org/document/10159843
Additional URL:	https://www.researchgate.net/publication/372020426...
Document Type:	Conference or workshop publication
Year of publication:	2023
Book title:	2023 46th MIPRO ICT and Electronics Convention (MIPRO) : May 22-May 26, 2023, Opatija, Croatia : proceedings
Page range:	56-61
Conference title:	MIPRO 2023
Location of the conference venue:	Opatija, Croatia
Date of the conference:	22.- 26.05.2023
Publisher:	Čišić, Dragan ; Vrcek, Nerven ; Koricic, Marko ; Gradisnik, Vera ; Skala, Karolj ; Čičin-Šain, Marina ; Babic, Snjezana ; Sruk, Vlado ; Skvorc, Dejan ; Jovic, Alan ; Gros, Stjepan ; Vrdoljak, Boris ; Tijan, Edvard ; Katulic, Tihomir ; Petrovic, Juraj ; Galinac Grbac, Tihana ; Bozicevic, Lovro
Place of publication:	Opatija, Croatia
Publishing house:	IEEE
ISBN:	978-1-6654-9420-5 , 978-953-233-104-2 , 978-953-233-105-9
ISSN:	2623-8764
Publication language:	English
Institution:	Business School > Enterprise Systems (Höhle 2017-)
Subject:	330 Economics
Keywords (English):	open , source , software , developer , analytics , dashboard , contribution , length , sentiment
Abstract:	In open source software development it is essential to stimulate developers’ contribution behavior in the absence of monetary incentives. But research on potential secondary effects of these stimuli, such as on contribution sentiment, remains scarce. Thus, we investigate how a promising external stimulus, an analytics dashboard, influences developers’ number of contributions as well as associated message length and sentiment. Building on social comparison and goal setting theory we hypothesize that adopting the analytics dashboard increases developers’ number of contributions but decreases contribution message length and sentiment. To test these hypotheses, we leverage granular data from a matched sample of 43,434 GitHub developers over two years who adopted an analytics dashboard on their personal developer page. Our difference-in-differences analyses reveal that adopting the analytics dashboard increases a developers’ number of contributions and associated message length. However, after adopting the analytics dashboard users write more negative messages and this effect is primarily driven by users with low initial GitHub activity. These findings suggest that the performance increase through analytics dashboards comes at the cost of potentially harmful consequences for the developers or the platform. We discuss implications for theory and practice.