Balancing Rewards and Behaviors: A Case Study

A Three Minute Read

A few years ago, an audio recording of the “Customer Service Rep from Hell” was the darling of the internet. A shrieking, belligerent agent refused, for over 15 minutes, to allow a caller to cancel his Comcast service. Mea culpas aplenty were issued, but each statement by Comcast pointedly through the rep under the bus; if there was organizational accountability taken, it hasn’t shown up on Google yet.

One such “apology”:

The way in which our representative communicated with [the (former) customer] is unacceptable and not consistent with how we train our customer service representatives… we are using this very unfortunate experience to reinforce how important it is to always treat our customers with the utmost respect.


Applying What I Learned from George Carlin

Sure, that may not be how they train their reps, but it is certainly what they both expect and accept. Agents are measured on customer retention – if they “allow” someone to cancel, they are grilled by their supervisors. If they convince a caller to not cancel service, they are praised. Inasmuch as shrieking at customers is “unacceptable”, allowing them to take their business elsewhere is even more acceptable.

Yelling at a customer service rep is as effective as yelling at your breakfast cereal, and it certainly is no punishment.

The calculus: Risk getting snapped at by a remote voice on the phone (likely never encountered again), or risk getting on the boss’s bad side – or worse? Hell, that agent probably gets a high five when he gets screamed at and still retains the customer. It’s worth it for the agent because that’s the way the system is rigged.

Actions speak louder than words – regardless of how much an organization talks about priorities (e.g., customer service), for the individual employee it comes down to WIFM: What’s In It for Me?

Using the Variables to Map Current State

We want to pinpoint both the desired and the undesired behaviors, and start thinking about how we want the environment to reward (or repel) those behaviors.

The desired behavior is that the agent convinces the customer to maintain his subscription. The undesired behavior is that the agent lets the subscriber cancel. (Note here that the “desired” and “undesired” are not simply opposites of each other, that is “he does one thing”/”he doesn’t do it”).

  • What encourages him to keep the customer on, as long as it takes?
    The avoidance of surely being singled out for letting a customer go.
    The chance of a small bonus.
  • What discourages him from keeping the customer on, as long as it takes?
    Having to deal with an angry public.
  • What will encourage him to simply cancel the service?
    The fear of potentially being yelled at by the caller.
  • What will discourage him to simply cancel the service?
    The more likely chance of being counseled for not putting forth effort.

One thing to capitalize on – this exercise should call into question what the desired behavior actually is. In this Comcast example, it’s a selfish, short-term success for the organization – and in conflict with any customer-centric behaviors. Hopefully going through this should spark some discussion about what Comcast really expects from its consumer-facing staff.

No matter what, once a caller identifies himself as someone thinking about cancelling service, the Agent knows there will be some level of pain – although there’s a small chance he will receive some money. When we think of the certainty of getting grief from his supervisor – whom he sees every day, and sits in judgment of his payday – versus the possibility of agitating the caller – who he never actually sees, and will likely never speak to again – there clearly is no other choice. Do whatever it takes to prevent the caller from cancelling.

To correct this situation, the behaviors don’t necessarily change, but managers need to think about what they can do to positively reinforce the desired outcome – one that works for both employee and customer.

Photo (c)Giu Vicente