Dan Heath

Mar 3 · 7 min read

In 2012, Ryan O’Neill, the head of the customer experience group for the travel website Expedia, had been sifting through some data from the company’s call center. One number he uncovered was so far-fetched as to be almost unbelievable. For every 100 customers who booked travel on Expedia — reserving flights or hotel rooms or rental cars — 58 of them placed a call afterward for help.

The primary appeal of an online travel site, of course, is self-service. No calls necessary. Imagine a gas station that allowed you to swipe a credit card right at the pump — and then, about 60% of the time, something went wrong that forced you to go inside the store for help. That was Expedia.

Traditionally, the call center had been managed for efficiency and customer satisfaction. Reps were trained to make the customer happy — as quickly as possible. Short calls minimized expenses. “The lens we were using was cost,” said O’Neill. “We had been trying to reduce that cost. Instead of a 10-minute call, could we make it a two-minute call? But the real question was: Why two minutes? Why any minutes?”

When you spend years responding to problems, you can sometimes overlook the fact that you could be preventing them. O’Neill shared his findings with his boss, Tucker Moodey, the executive vice president of global customer operations. Together, they dug into a basic but neglected question: Why in the world are so many customers calling us? They compiled a ranking of the top reasons customers sought support.

A “war room” was assembled, where people from different operating groups met on a daily basis, and the group was given a simple mandate: Save customers from needing to call us.

The number one reason customers called? To get a copy of their itinerary. In 2012, roughly 20 million calls were logged for that purpose. Twenty million calls! That’s like everyone in Florida calling Expedia in one year.

At a support cost of roughly $5 per call, that’s a $100 million problem. So why weren’t customers receiving their itineraries automatically? The answers were pretty simple: The customer had mistyped her email address. Or the itinerary ended up in her spam folder. Or she deleted the itinerary by accident, thinking it was a solicitation. Compounding the problem was that there was no way on the website for customers to retrieve their itineraries.

O’Neill and Moodey took their data to Dara Khosrowshahi, then the CEO of Expedia. “We’ve got to do something about this,” O’Neill recalled saying. Khosrowshahi not only agreed with their focus on reducing call volume, he made it the customer experience team’s top priority. A “war room” was assembled, where people from different operating groups met on a daily basis, and the group was given a simple mandate: Save customers from needing to call us.

The war room group deployed solutions for the top drivers of customer calls, knocking off one at a time. The fixes for the number one issue — the itinerary requests — came relatively quickly: adding an automated option to the company’s voice-response system (“Press two to resend your itinerary”), changing how emails were sent to avoid spam filters, and creating an online tool to allow customers to handle the task themselves.

Since 2012, the percentage of Expedia customers who call for support has declined from 58% to roughly 15%.

Today, virtually all of those calls have been eliminated. Twenty million support calls just vanished. Similar progress was made on the other “top 10” issues. Since 2012, the percentage of Expedia customers who call for support has declined from 58% to roughly 15%.

The effort to reduce call volume at Expedia was a successful upstream intervention. Downstream actions react to problems once they’ve occurred. Upstream efforts aim to prevent those problems from happening. You can answer a customer’s call and address her complaint about a missing itinerary (downstream), or you can render that call unnecessary by ensuring that she receives her itinerary up front (upstream).

Surely we’d all prefer to live in the upstream world where problems are prevented rather than reacted to. What holds us back? Looking back on Expedia’s success, what’s particularly hard to understand is why it took so long to act. How could the company have reached the point where 20 million people were calling for itineraries? Shouldn’t the alarm bells have been ringing rather loudly by the time, say, the 7 millionth call was logged?

Expedia’s executives were not oblivious. They were aware of the huge volume of calls. It’s just that they were organized to neglect their awareness. Like most companies, Expedia divided its workforce into groups, each with its own focus. The marketing team attracted customers to the site. The product team nudged customers to complete a reservation. The tech group kept the website’s features humming along smoothly. And the support group addressed customers’ issues quickly and satisfactorily.

Notice what was missing: It was no group’s job to ensure that customers didn’t need to call for support. In fact, no team really stood to gain if customers stopped calling. It wasn’t what they were measured on.

In some ways, the goals of the groups actually encouraged more calls. For the product group, whose goal was to maximize bookings, the best move was to ask for a customer’s email only once, because asking her to type it a second time would add friction. They might lose one person in 100 who’d be annoyed enough to abandon the transaction.

But the side effect of that decision, of course, is that some customers would mistype their emails, and they’d end up calling for an itinerary. That’s a system failure. That customer never needed to call. Yet both teams would still look like heroes according to their goals: The product team closed a transaction, and the support team handled the resulting call quickly.

The specialization inherent to organizations creates great efficiencies. But it also deters efforts to integrate in new, advantageous ways.

Mark Okerstrom, who was Expedia’s CFO in 2012 and became CEO in 2017, said, “When we create organizations, we’re doing it to give people focus. We’re essentially giving them a license to be myopic. We’re saying: This is your problem. Define your mission and create your strategy and align your resources to solve that problem. And you have the divine right to ignore all of the other stuff that doesn’t align with that.”

Okerstrom’s point is that focus is both the strength and the weakness of organizations. The specialization inherent to organizations creates great efficiencies. But it also deters efforts to integrate in new, advantageous ways. In upstream ways.

And this is true in many parts of society. So often in life, we get stuck in a cycle of response. We put out fires. We deal with emergencies. We handle one problem after another, but we never get around to fixing the systems that caused the problems.

Therapists rehabilitate people addicted to drugs, and corporate recruiters replace talented executives who leave, and pediatricians prescribe inhalers to kids with breathing problems. And obviously it’s great that there are professionals who can address these problems, but wouldn’t it be better if the addicts never tried drugs, and the executives were happy to stay put, and the kids never got asthma? So why do our efforts skew so heavily toward reaction rather than prevention?

Back in 2009, I spoke with a deputy chief of police in a Canadian city; it was one of the conversations that sparked my interest in upstream thinking. He believed that the police force was unduly focused on reacting to crimes as opposed to preventing them. “A lot of people on the force want to play cops and robbers,” he said. “It’s much easier to say ‘I arrested this guy’ than to say ‘I spent some time talking to this wayward kid.’”

He gave an example of two police officers. The first officer spends half a shift standing on a street corner where many accidents happen; her visible presence makes drivers more careful and might prevent collisions. The second officer hides around the corner, nabbing cars for prohibited-turn violations. It’s the first officer who did more to help public safety, said the deputy chief, but it’s the second officer who will be rewarded, because she has a stack full of tickets to show for her efforts.

That’s one reason why we tend to favor reaction — because it’s more tangible. Downstream work is easier to see. Easier to measure. There is a maddening ambiguity about upstream efforts. One day, there’s a family that does not get into a car accident because a police officer’s presence made them incrementally more cautious. That family has no idea what didn’t happen, and neither does the officer. How do you prove what did not happen? Your only hope, as a police chief, is to keep such good evidence of crashes that you can detect success when the numbers start falling. But even if you feel confident your efforts accomplished something, you’ll still never know who you helped. You’ll just see some numbers decline on a page. Your victories are stories written in data, starring invisible heroes who save invisible victims.

I prefer the word upstream to preventive or proactive because I like the way the stream metaphor prods us to expand our thinking about solutions. For any problem, we can intervene at many points along an almost limitless timeline. In other words, you don’t head Upstream, as in a specific destination. You head upstream, as in a direction. If you’re trying to prevent kids from drowning, then swim lessons are further upstream than life preservers. And there’s always a way to push further upstream — at the cost of more complexity.

It’s not that the upstream solution is always right. And it’s certainly not the case that we should abandon downstream work — we will always want someone there to rescue us. The point is that our attention is grossly asymmetrical. We spend our lives consumed with problems that we might have avoided altogether.