Managing employees who work together globally is an ongoing challenge for executives of large organizations. The ability to quickly detect changes in the health of relationships with clients can unlock significant value. Many companies monitor customer satisfaction through customer satisfaction surveys, such as the Net Promoter Score (NPS) pioneered by Bain & Co. However, such methods are limited in their ability to provide frequent, detailed and cost-effective evaluations.
The way people interact with each other and what they say offer an important window into how they feel about each other. However, the content of what is said is often less significant than how it is said and the accompanying body language. Although face-to-face meetings between providers and customers may offer the clearest and most comprehensive way for managers to gauge customer satisfaction, such information is not generally available in high-volume, globalized collaborations.
The authors have developed an assessment method for analyzing email communication patterns between customers and vendors in geographically distributed environments where face-to-face meetings are impractical or impossible. In conducting the research, the authors analyzed the email interactions of 32 service delivery organizations — all part of Genpact, a large global services provider. After notifying the various parties involved and obtaining permissions, the authors retrieved the communications between account managers and their customers and constructed the social network.
The authors collected the email boxes of two to three account executives per customer studied and compared the network structure of the communications to the customer’s NPS, which was the primary customer satisfaction metric Genpact had been using. The authors also collected the email headers from incoming and outgoing emails for the account executives over a period of eight months. The researchers compared the structure of the resulting communication network with the NPS scores.
The authors acknowledge that the method is specifically based on email analysis, and thus it is poorly suited to work contexts where email is not widely employed. In the future, they suggest that it might be possible to adapt the methodology to other electronic media, such as instant messaging services and voice-over-IP telephony.