Manage Learn to apply best practices and optimize your operations.

Expedia's focus on customer intelligence turns clicks into dollars

Expedia drills into its data and finds four ways to better serve its customers and build a stronger site with customer intelligence data.

Just a week into his new job at Expedia Inc., Joe Megibow started to notice a pattern. As he was studying the company's operations, he was met with the same, persistent inquiry from employees.

"It was a well-formed and eloquent question," said Megibow, vice president and general manager of the online travel site "It was, 'Why does our checkout conversion suck.'"

That question turned out to be a good one, illuminating another, harder-to-see pattern: While website traffic was healthy, numbers fell off during the last step of the process to book a flight or a hotel room. For Megibow, something just didn't add up and begged the question: "Why [were] they going this far into the process and not booking?"

The answer turned out to be simple and unexpected. For some customers, the process failed as they pressed the purchase button. Megibow, who pulled an all-nighter combing through data to pinpoint the problem, recounted the details of this story during his session "Find Your Creative Customer Analysis" last month at SAS Institute Inc.'s Premier Business Leadership Series conference.

His presentation was decidedly unique. Unlike many presenters who tout their successes, Megibow provided lessons learned from failures like this one from six years ago. Collectively, his stories showcased how Expedia has grown over the years, building to, arguably, his most important lesson: It's not just about data, reports and analysis. The real value comes from action. That, Megibow said, means making decisions, testing those decisions and, yes, making mistakes.

"We are wired as human beings to fail to learn," he said. "You've got to make those mistakes."

His lessons have been broken into a two-part series. This first installment focuses on the practice of customer intelligence and data; the second on building a data strategy.

Lesson 1: It's not just about click conversions, it's also about intent

It turned out visitors who attempted to open a new account with an email address already on file ran into difficulty -- especially if the new and existing passwords didn't match. Those customers were met with an error message telling them their passwords were invalid, but only after providing the necessary information to make the purchase. To make matters worse, customers weren't told why the password failed. Eventually, they abandoned their purchases.

What is customer intelligence?

Customer intelligence (CI) is information derived from customer data that an organization collects from both internal and external sources. The purpose of CI is to understand customer motivations better in order to drive future growth. The application of business analytics to customer data is sometimes called customer data mining.

For more information, read the definition of CI.

Studying conversion data wouldn't have revealed the problem; Megibow uncovered it only when he sought to understand customer intent. The failure to connect customers with potential purchases was certainly a problem, but the more alarming detail was the frequency in which it occurred.

"It was happening to like one in five people who were clicking on the 'take my money' button, and it wasn't working," Megibow said.

Lesson 2: Translate data into actionable terms

Although the site was completely operational, Megibow took the news to his boss and explained they were experiencing an outage. That label was not used accidentally; after walking his boss through data that showed almost 20% of customers who wanted to make a purchase couldn't, Megibow made the comparison concrete: "You might as well shut the site down for five hours a day," he said.

The failure of the purchase button equated to the business turning its back on a sale over and over again. To get his boss on board, he needed to express the seriousness of the situation. "It was translating data into a very consumable, actionable term," he said.

And convince his boss he did. Following the typical outage protocol, the IT troops were rallied and all other projects were put on hold until this problem was fixed, which happened in a matter of days. Almost overnight, Megibow said, Expedia saw a dramatic increase in booking confirmations.

Lesson 3: If you give your customers a voice, they will give you gold

Pinpointing and solving this problem helped broaden Expedia's interest in learning about customer intent. But it also shed light on something else: Having Megibow sift through data and pull all-nighters was not scalable.

"It was like looking for needles in a haystack," said Megibow. "There had to be a better way."

More on Expedia

Read about Expedia's quest for excellence in IT processes

Find out why Expedia snubbed hosted tools for machine-generated data management

Learn why being sticky has worked for Expedia's UK site

Megibow and his team took stock of Expedia's data and realized other channels could prove to be valuable -- specifically, data generated by feedback loops. These sources were used for overall trend analysis and little else, but combining the two -- the user experience data, which they had, with the actual language from feedback data, such as website comments -- could reveal intent.

"This was my Reese's Peanut Butter [cup], you've-got-your-chocolate-in-my-peanut-butter moment," he said.

Soon, customer feedback was taking on more importance. Megibow and his team even began serving up this combined data in daily emails to executives, which became popular -- and useful. For example, when a highly-valued customer complained he couldn't access his account, the complaint did not go unnoticed. Instead, it led Megibow to realize more work was yet to be done.

"It turned out, in fact, that we were literally deleting from our system thousands of accounts every single day," Megibow said. "And the funny part is, until there was an economic value associated with it, we never heard about it."

Lesson 4: Less is almost always more

Megibow and his team continued digging into customer feedback, looking for other places that needed attention or refinement. They landed on the payment information form.

At first glance, the payment information page looked like any other: A single Web page with fields for information -- credit card number, name and billing address. But a closer look revealed a subtle difference: After the customer enters his name and credit card number, the next empty field encountered asks for "company name."

"It just kind of throws you," Megibow said. "Sure, it says optional, but it's just right in the flow."

Customers appeared to have trouble with this particular field. The data showed that confused customers, with credit cards still in hand, guessed Expedia wanted more credit card information, and so typed in "Chase" or "Bank of America." When the next field asked for an address, those still-confused customers filled in the address for the lending agency rather than a billing address.

The field for "company name" was eventually deleted. Doing so drove $1 million in profits a month.

"There is a temptation of analysts and marketers to say, 'Oh, if you just gave me a little more information, I can give you a much better experience,'" Megibow said. "But this stuff can create friction."

Instead, help customers easily move through the process, provide them with a good experience, and businesses will develop longer-term relationships. The data, in other words, will come eventually.

Dig Deeper on Customer analytics

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.