Digitizing High-Touch Interactions : The Divided Efficacy in Personal Selling

We study how digitization affects the efficacy of high-touch interactions, specifically by looking at the relative value of in-person versus virtual interactions (e.g., WebEx, Facetime). In a direct sales context, our empirical strategy exploits a pseudo-random experiment performed on thousands of clients overseen by agents of a large medical device firm. A subset of agents switched from entirely using in-person visits (IPV) to a mix of remote visits (RV) and IPV. By studying the contemporaneous and spillover effects of RV and IPV, we find that, on average, IPV represents nearly one-and-a-half times the value of RV. We leverage advances in machine learning to explore if there are factors that may shrink or magnify this gap. Clients who have stronger ties to the focal firm are relatively less responsive to RV, which widens the gap. If a client is in a technology-prone region, we find the gap shrinks to the point that the effect of RV is near parity with IPV