Customer touchpoints provide valuable data to customers shopping for products. Is this a data source for marketing? Would it be possible for us to capture call center conversations, web chats, email trails, and so forth to gauge customer interest? Can we also use this data to understand differences across customer cross sections?
Marketers are increasingly using customer touchpoints to communicate with customers, and a fair amount of this data is already being captured for further analysis. A cable provider recently asked me to analyze their call center data to ascertain customer intention. In their environment, the call center agents were codifying customer intent using 150+ codes at the end of the call. As suspected, the reported information was not necessarily accurate. Most of the call center agents used fewer than ten codes to represent most of their conversations. It was not clear whether customers were calling only for those reasons or whether the call center agents could memorize only a small number of codes and were repeatedly using those codes to inaccurately represent the real conversation.
However, a mechanical means of data categorization and mining can be used to verify and autocorrect these observations. Call center conversations can be converted into text, and then the text can be categorized into a code. In addition, the customer voice analysis can detect anger or appreciation, leading to not one but possibly many quantifications of the call center information. Web chat and email can also be analyzed, and although they lose the verbal emotions, they can be used for written emotions, for example, use of adjectives and adverbs to codify emphasis and emotions.
The most powerful analysis from conversation data is in its use to identify gaps between intention and action. If someone is interested in purchasing a product, but does not end up buying in a specific contact point, it represents an unfulfilled demand that can be further addressed via a campaign.