Welcome to the Interaction Metrics Blog
Back to Insights

This Sunday in The Haggler (David Segal. “In a Stock Trade, It’s Man vs. Machine.” New York Times, April 27, 2014), David Segal looked at a situation in which a computer misunderstood human language. We’re not surprised. As we’ve said before, customer experiences are complex, so computers don’t always get it.The-Haggler-NYT

In this particular case, a customer intended to use the IVR system at E*TRADE to sell 5 shares of his Apple stock—but the computer sold all 119 shares in his account. When the customer contacted E*TRADE to resolve the issue, they responded with an ambiguous, stock response:

“A trade inquiry was completed on Sept. 10, 2012, and it was confirmed that the order was properly handled per your instructions.”

What was this “trade inquiry”? One would think this meant that a thorough customer service evaluation was performed—that an auditor listened to a recording of the customer’s instructions to confirm the computer got it right. But as David Segal learned, E*TRADE didn’t even have a recording available.

OUR TAKE:

The customer service technologies that E*TRADE uses are essential. They keep costs low for companies and make routine transactions efficient for customers. However, E*TRADE (and plenty of other companies) are over-trusting of their software, especially when it comes to customer service.

The E*TRADE system is set up as though the machine is infallible and customer service evaluation unnecessary. As this instance shows, customer service evaluation is essential—and that’s especially true when things go awry. What E*TRADE needed was human process to identify the error and provide the customer with a meaningful response. They didn’t have this process—so they wound up feeding David Segal a quote, ripened and ready for Sunday’s Haggler.

Let's Discuss Customer Listening:

Don't forget your name
Don't forget your company
Required format: bob@example.com
Click here if you're not a robot.

Don't forget to check the box above. It helps us prevent spam.