Title: Attribution modelling in digital advertising for e-commerce

Authors: Sergey Alexandrovskiy; Olga Trundova

Addresses: National Research University Higher School of Economics, Office 405, Rodionova Str 136, Nizhny Novgorod, 603039, Russia ' National Research University Higher School of Economics, Russia

Abstract: The authors aim to offer a probability approach for measuring media contribution to online conversions in e-commerce. The authors reviewed literature on attribution modelling with application of heuristics (Google Analytics) and probability (Markov chains) models. The survey used the data of customer journeys from 134,132 users to build up the attribution model. As a result, Markov chains model minimises the value of direct online channel, and redistributes the value in favour of other traffic channels and gives a significant weight to the paid traffic. Markov chains model increases the importance of display and paid search channels, which is not so noticeable in heuristic models. Findings of this study help marketers to apply Markov chains modelling in online advertising evaluation and budgeting. The authors proposed application of Markov chains attribution modelling in the e-commerce website. Other e-commerce companies might apply Markov chains attribution modelling in a similar way.

Keywords: attribution modelling; digital advertising; e-commerce; Markov chains model; multi-channel measurement; probability model.

DOI: 10.1504/IJIMA.2022.120964

International Journal of Internet Marketing and Advertising, 2022 Vol.16 No.1/2, pp.19 - 37

Received: 06 Oct 2019
Accepted: 03 Mar 2020

Published online: 21 Feb 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article