In customer service, things are bound to go wrong, at some point. When that happens, it is important to not only understand what went wrong, but also what is the perceived cause of the problem, because that impacts on the recovery strategy.
Take interacting with a chatbot, for instance. As discussed in a previous post, many factors can lead a customer to be unhappy after interacting with a customer service chatbot. Here are the 5 categories of problems that can occur:
|Functionality||Chatbot is deemed to be of limited assistance|
|Affective||Chatbot lacks empathy|
|Integration||Loss of information during interaction, or during handover to human assistant|
|Cognition||Chatbot can’t understand query|
|Authenticity||Unclear whether service is being provided by chatbot or human|
To properly address customers’ frustration caused by these problems, we need to not only solve the real cause of the problem, but also the perceived cause. Why? Because the perceived cause will shape customers’ response to the problem, as well how they interpret other information from or about the firm. We call this effect, the confirmation bias. For instance, if customers believe that the problem occurred because they are unable to use chatbots, they will use an alternative channel to contact the firm or, maybe, try to become more “versed” at using this technology. However, if they feel that it was the firm’s fault, they will seek redress (e.g., an apology, or compensation) or, even, retribution.
So, who do customers blame, when things go wrong in interactions with chatbots?
According to research reported in the paper co-authored with Daniela Castillo and Emanuel Said (mentioned in a previous post), these are the perceived sources of problems when interacting with customer service chatbots:
|Problem||Who’s to blame?||Firm, because:||Customer, because:|
|Functionality||Firm, only||Failure to invest in features valued by customer [care]||—|
|Affective||Firm, mostly||Failure to invest; Unsuitable use of chatbot [care]||Diminished presence of mind, during emotional situations [emotional]|
|Integration||Firm, mostly||Poor organizational procedures [incompetence]||Failure to proactively save information [behaviour]|
|Cognition||Both||Lack of attention during development of the chatbot [incompetence]||Failure to use simple language and check spelling [cognitive]|
|Authenticity||Both||Deliberate misrepresentation of chatbot via naming, script and other features [deceit]||Failure to notice cues – e.g., speed of reply [cognitive]|
We found that, in situations of “mild” dissatisfaction, customers tried to by-pass the chatbot, and solve the problem via another channel, which results in resource inefficiency for the firm. This would include resorting to social media to express their frustration, which can impact how other customers perceive the brand. However, when customers experienced more dissatisfaction and, specially when they experienced emotional losses (e.g., affective problems), they pursued harsher measures such as terminating the service, or moving to a competitor, which represents loss of revenue for the firms.
So, when measuring the success of a chatbot, don’t think only in terms of “cost savings” or “convenience”. Think also how it can impact the firm’s image, and how customers’ reason about the causes of the problems that they experience.
The paper reporting this study was published in The Service Industries Journal. You can find a free, pre-print version of the paper,here.