The Data Whisperer

Marta Rodriguez Martinez is trained as an industrial engineer. "I studied industrial engineering because I was fascinated by energy and power generation. I imagined myself in the middle of a power plant fixing things," said the Microsoft travel manager for global travel business intelligence and analytics. "Who knew I would be working with travel data?"

Whether she thinks so or not, her colleagues at Microsoft believe she is indeed fixing a lot of their problems and, even more, facilitating overdue innovation for 75,000 Microsoft travelers.

Since last year, Microsoft has worked to overhaul its travel program. The company has outsourced day-to-day travel management operations to its agency, American Express Global Business Travel, but drives strategic direction internally, making sure it owns and analyzes all its travel-related data to advance that effort.

Director of travel, venue sourcing and payment Eric Bailey largely has ditched travel policy, opting instead for a traveler-centric approach that prioritizes listening—that is, listening closely to travel data to understand where, why, how, how often and with which suppliers Microsoft employees travel. Based on this listening campaign, Microsoft now divides its travelers into four persona groups. "We can't know all of our travelers," said Bailey. "What we can do is take their data and turn it back into travelers, or at least types of travelers."

Rodriguez Martinez is at the center of that effort. She manages and manipulates 11 data streams that flow into Microsoft's data lake. She uses Microsoft's Azure Machine Learning Studio and its Power BI to develop insights into traveler booking and spend behavior, to analyze loyalty to suppliers and to determine whether travel patterns are fixed or can be influenced by better supplier choices or added value.

Letting machines do the number crunching has made the travel team "very efficient listeners," Rodriguez Martinez said. "We have a lot of data streams that we can integrate, and we can really get to the key messages that travelers are throwing to us" through booking data, credit card and expense data but also through unstructured platforms like social feeds from Yammer, traveler surveys and customer relationship management platforms.

Future Vision & What's Happening Now

Bailey's vision of predictive booking and frictionless travel experiences defines a lot of the machine learning and artificial intelligence efforts happening on the front end of travel management at Microsoft. "The simplest way to look at [what we're trying to do with machine learning and AI] is to remember what a real conversation with a good travel agent used to be like," he said.

Rodriguez Martinez picked up the thread: "Imagine if you had a machine that was intelligent enough to know when you're traveling, all your preferences, all your passions and prompted you to book because the machine saw a meeting on the calendar: 'I can see you're traveling. You normally take this flight. You can make it on time for the meeting. This is the price. Do you want me to book it for you—yes or no?' And then you are done."

Rodriguez Martinez acknowledged that such a frictionless front-end experience is a longer-term goal. "But there are a lot of corporate use cases for machine learning and AI when it comes to strategic decision-making. We can do that right now."

She delivers deep data intelligence to Microsoft category managers to inform sourcing and contracting discussions. Armed for the first time with persona data, Microsoft global travel sourcing manager Diane Lundeen Smith will initiate her airline RFP process this summer, but she already has leveraged deeper intelligence in day-to-day supplier management. "With Marta's help, I look at scorecards to track [program] performance in different areas: standard cost savings, account management and procurement measurements," Smith said. Contract utilization is a new report that allows Smith to look at how often the Microsoft deals she negotiates are actually used by travelers.

"You can look at a [general] marketshare picture for a partner airline, for example, and not know how much of that was [the Microsoft] deal versus a published fare, Web fare, agency fare or some other rate," Smith said. With machines crunching the numbers and comparing all transactions on that airline against contracted rates, Smith gets clarity around the value of her work with that partner.

The next step is to get details: Why isn't the contract being used or what has changed with the supplier or with the business need? This is where Rodriguez Martinez's work with unstructured data can come into play. She consolidates feedback from Microsoft's internal social feed, standard travel surveys and CRM data to create a picture of what Microsoft travelers think about suppliers. Rodriguez Martinez said she can look at this voice of the traveler data to uncover traveler concerns. Plus, she said,  "they drop a lot of hints about their supplier loyalty."

Smith used this unstructured data recently to settle what she called "a dust up" with Emirates in October. "They had made changes with pre-assigned seating and the amount of baggage you can bring. Marta's Voice of the Traveler data enabled me to approach Emirates with traveler verbatims and clearly articulate the issues we were facing." That Smith was able to show the travelers were valuable customers to Emirates drove her case home. "I was able to recommend alternatives that would benefit our travelers and negotiate a solution that was particularly important to those who travel extensively in Emirates' markets."

Developing Data Insights

Rodriguez Martinez hails from travel data intelligence firm Pi, which cleanses, normalizes and crunches travel data for corporate clients. The experience honed her deep data skills before her arrival at Microsoft. Even so, she said better tools make it easier for travel managers and other non-expert users to facilitate deeper data analysis. "As a travel manager, you [need to] understand the potential of machine learning for your own benefit," she said. "People need to embrace it instead of saying, 'Oh, you're doing that because you're Microsoft,' or, 'You're doing that and that's really cool, but I don't know anything about AI or machine learning and I'm too scared to dig in.'"

At Microsoft, developing new data insights is a collaborative effort, and getting from questions to answers isn't always a clean process. "We brainstorm a lot and we come up with more ideas and we sidetrack a lot but then carry on," Rodriguez Martinez said.

Smith underscored the sense of discovery. "Those are my favorite sessions: when we are running different queries and I say, 'Wait, can it do this?' I don't know everything that is possible with the data, so I often don't know where we are going to end up."

Yet Rodriguez Martinez is careful to steer clear of the theoretical. "Data for the sake of data is boring; it's just a headache," she said. She keeps the team grounded in what she calls "user stories" to ensure Microsoft's data efforts focus on actual challenges and deliver meaningful results. "I need to know what questions they want to ask the data and what actions will be taken [once we get] answers. You have to start with the thought process behind the request and stay focused on how the data can be used," she said.

Rodriguez Martinez, Smith and Microsoft global travel experience lead Julia Fidler recently collaborated to evaluate how travelers in different countries perceive the negotiated rates offered through Microsoft's managed channels. Fidler and Smith wanted to understand, per country, if better rates were available on the market or if the program suffered from pockets of perception issues. The answers would determine whether the problem required a sourcing solution or an outreach and communications solution.

"I took [market] pricing data and [agency] booking data and combined it with CRM data and some [additional] sources," said Rodriguez Martinez. "We found, for example, that price confidence was very low among travelers in Brazil. When we ran tests on the actual pricing data, though, we didn't see the price issues." Ultimately, she said, Brazil suffered from a price perception issue. "Diane now knows that it's not her issue—it's Julia's issue—and we know how to take action."

Even with clear direction, there's creativity in play, and the data-discovery process benefits from multiple viewpoints and layering ideas. Rodriguez Martinez admits she's not a sourcing expert, and yet Smith recognizes the value in a fresh approach. "Marta does the magic. She is able to take what we are looking for and come up with ideas of how to get answers. Because she has been outside of traditional travel roles, she brings a different viewpoint and she allows us to push the envelope and to evolve our program toward the future."

Driving the Industry Forward

While Bailey's vision of the front-end traveler experience is Microsoft's priority, Rodriguez Martinez suggested that procurement innovation is the biggest opportunity for travel data in the near term. And, ultimately, the two can't be separated.

IATA's New Distribution Capability and other distribution strategies, for example, have the potential to tie front-end and back-end data together in new ways. They provide the ability to tailor travel product offerings to or to bundle benefits for specific corporations or, in Microsoft's case, to different persona groups within the organization. That means that the more a corporation knows about its travelers, the better products and benefits it can deliver to them through managed tools. The more granular the detail in that data, the more predictive and relevant those offers and product bundles can be.

These possibilities are changing the way Microsoft approaches travel management and supplier management. "We have to develop programs in a way that enables people to purchase travel in new ways," Rodriguez Martinez said. "The data is not just about taking; it's also about giving information to suppliers. That's how I see the way forward. In terms of machine learning and how we understand our own travelers, suppliers can learn from what we have to offer, and we can learn from what they offer."

Smith has already started to change her conversations with suppliers based on the data that she can provide and her expectations for future innovation. "When you think about managing a program Microsoft's size, it's been challenging to get that deep with suppliers. With these numbers, you can start to see the evolution and how we can get into deeper and more innovative partnerships," said Smith. "We may have fewer partners that are more innovative, that can tailor offers that meet the needs of our travelers and then can get those offers to the travelers at the point of sale. Our suppliers are super interested in exploring this with us."

In the meantime, Rodriguez Martinez continues the effort to understand travelers through their data. "We are still building the algorithm of where it can go, how it can serve our travelers better and how it will inform our program," she said. "We have lots of thoughts about it: how to differentiate who is predictable and who is not and how we can influence travelers through better, more relevant options. Because even if you have flown one carrier forever, are you doing that because you are loyal or because there are few options? If your company could provide a better offer, would you take it? Probably."