Imagine you just got home after a long day at work and you sit on your couch to watch Netflix. But to your horror, the modem lights are blinking and you can’t access your account online. Frantically, you pick up the phone to call your Internet provider and you hear: “In a few words, please tell me what you’re calling about.”
At first glance this seems like an easy task because all you need to do is explain the situation. But how do you explain the problem if you don’t know what the problem is?
In many instances of customers calling for technical support, customers are reaching out because they don’t know what the problem is. In other situations, callers are extremely familiar with their issue and just need a quick fix.
Whether you’re a technical newbie or an experienced troubleshooter, you need to find an accurate and concise way to describe your problem and you need to do it in just a few seconds.
Callers who have experienced this situation before – and know that the internet connection is down when the modem lights are blinking – may find this task easier than others. They may say something like, “My internet is down,” or “My connection isn’t working.” Callers who haven’t experienced this may find the task more daunting. They may respond with something like “Uhh I think there’s something wrong with my uhh…modem.” Callers who are unfamiliar with this problem may benefit from more direction. For example, if they were greeted with: “In a few words, please tell me what you’re calling about. You can say things like ‘My internet is down’ or ‘Pay my bill’.”
We at Nuance wanted to understand caller behavior when determining the reason for their call, so we set up an experiment. We asked callers for their intent in two different ways: 1) No examples: “In a few words, please tell me what you’re calling about” and 2) With two examples: “In a few words, please tell me what you’re calling about. You can say things like ‘My internet is down’ or ‘Pay my bill’.” We deployed the experiment across multiple applications and different verticals.
We found that when there are no examples, callers provide a wider variety of phrases in response; they are less prone to ask for an agent, and they are less prone to transfer to the wrong agent or someone who can’t help. These findings suggest that callers enjoy more expressive freedom when they aren’t given examples or restrictions in terms of how to phrase the reason for their call. It’s possible that callers who hear examples feel limited in what they can say and ask for an agent as a result. Since generic agent requests usually transfer to a default queue, many callers need to be transferred to another agent.
However, we also found that callers are more prone to say “um” and “uh” when there are no examples. This is an indication they struggle to explain and find the task more difficult without examples of what to say or how to phrase an issue.
We have found variation in how we ask the caller for their intent can have a significant impact on their behavior. Supplying no examples may encourage expressive freedom, while increasing task difficulty. But keep in mind each situation is unique, so you should always test different strategies and weigh their pros and cons before moving ahead with one style or the other.
Tips for callers:
- Take a minute to think about how to express your question before picking up the phone.
- Jot down on a piece of paper a couple of keywords.
- While on the phone, don’t feel pressured to start speaking as soon as possible. The system will give you multiple chances to give your intent.
- If you really feel you can’t come up with the right words, start speaking and you’ll figure things out as you speak. Chances are you thought you didn’t have the right words but your actual utterance will sound coherent.
Tips for customer service providers:
- Test out different prompting strategies when collecting intent and evaluate the data. Choose the prompting strategy that works best for your caller population and for your business needs.
- If you have limited reporting capabilities, rotate prompting strategies.
- When providing examples, use phrases that unambiguously lead to self-service or to a specific agent queue.