The effect of a humanoid service robot versus a tablet service kiosk in retail: an observational study

Marketers are continuously searching for innovative ways to engage with customers. Humanoid service robots (HSR) are increasingly being used in retail and consequently changing the foundations of the servicescape. As the store environment is the main laboratory for consumer experiments (Larsen et al., 2017), this study jumps on the bandwagon of a scholarly growing interest in inductive in-store behavior research by means of unobtrusive observations. The present observational study also responds to a recent call by Wirtz et al. (2018) to examine customer-service robot interactions on a micro level-perspective. Additionally, it adds to the research of Aaltonen et al. (2017), by also observing initial responses to a HSR in a shopping environment, but more in depth, without intervening in the customer-robot interaction and over a more extended time period.

This paper is probably the first in-depth observational field study analyzing human-robot interactions in terms of real in-store and spending behavior. Furthermore, observed customer interactions with the HSR are systematically compared with customer interactions with a tablet service kiosk (TSK), to inspect what added value investing in HSR can offer to retailers. The research has been conducted in a branded chocolate store located at Brussels Airport. The analysis of the 42 hours of video-taped observational footage has been systematically performed along the hierarchy of the AIDA model (attention, interest, desire and action).

The experimental design of this field study consists of three conditions, namely a control, TSK and HSR condition. In total, six days (two days of seven hours per condition) were recorded, resulting in a total of 42 hours of camera footage. A quiz related to chocolate in general as well as to the specific products and brands inside the store was provided to the passengers via the two alternative media channels (TSK and HSR). Passersby were tracked when walking along the store and when starting an interaction with either the TSK or HSR. A total of 106,357 passengers (Ncontrol = 33,900, NTSK = 36,394, NHSR = 36,063) and 963 interactions were registered (NTSK = 34, NHSR = 929).

We expected the HSR to outperform the TSK at all four stages of the AIDA model and found this largely confirmed in our observations. With respect to the first stage of the AIDA model ‘Attention’, our research findings demonstrate that a HSR was able to attract more passengers to start an interaction, which also generated larger groups of bystanders (see video in De Gauquier et al., 2018). Regarding the second stage ‘Interest’, we found that while the HSR was able to trigger longer interactions, less people were actually looking at the store while interacting with the HSR as compared to the TSK. On the other hand, a longer interaction time with the HSR appeared to lead to higher odds of entering the store in comparison with the TSK, which indicates that their ‘Desire’ was triggered more via the HSR. Moreover, the HSR appeared to convince more people to enter the store in comparison with the TSK. Considering the last stage ‘Action’, however, relatively more people were recorded to have bought something after interacting with the TSK as compared to the HSR. Nonetheless, overall, a larger number of unique transactions was registered, and the average amount spent per customer appeared to be higher during the period when the HSR was present in comparison to the control and TSK conditions.

To conclude, this study provides retail practitioners with valuable insights to understand the merits of implementing a HSR. More specifically, the research findings prove that deploying a HSR as an in-store promotional instrument is more effective than a traditional TSK in terms of the AIDA model.