Should You Build a Data Tool Rather Than Buy?

Despite the myriad data management and reporting systems on the market, many travel programs develop their own solutions. This isn't a case of late adoption, however, as when Excel sheets stuck around even as automated expense reporting tools became more sophisticated and egalitarian. Rather, some companies are eschewing off-the-shelf tools for reasons of cost and/or functionality, assuming they can do a better job themselves. What's more, the barriers to entry for an in-house data management program are shrinking.

Depending on a travel buyer's background and the company's structure, managing data might fall outside his or her domain. Even so, many have marshalled the resources within their own companies and combined them with a few external building blocks to develop data systems that offer views into savings, compliance and supplier performance.

A self-built system isn't the right answer for all buyers, though, and so multiple suppliers are developing better tools that travel managers can buy to consolidate and interpret data.

Reasons to Build

When determining whether to buy or build a travel data tool, buyers should first outline what the company hopes to accomplish. From there, buyers will be better able to evaluate whether a tool that's already on the market could meet those goals in a cost-effective way.

Those who have opted to build their own solutions often express dissatisfaction with what's already out there. IBM travel council chair Bob Gisborne said travel management company reporting, for example, is not thorough enough. IBM asked its TMC to provide a travel data consolidation and analysis solution, but it was taking too much time, he said. IBM ran requests for information and proposals a few years ago and even identified one supplier that stood out. Ultimately, though, the company determined its money would be better spent developing a tool that could meet its needs exactly. "Our utopia was to put everything together that shoots out data to give us insights we can take action on, and it needs to be timely, real time if possible," Gisborne said. "We ended up realizing there wasn't anything out there, so we did it ourselves. I'd rather have something I want and know will work and spend those dollars working with our IBM research colleagues to develop something."

Similarly, Cardinal Health wanted data that works for a healthcare company and enables the messaging it needs to get out to stakeholders, head of indirect procurement Patrick Eckhert said. And Cardinal was not getting a clear picture of spending and key metrics from its TMC. "It was very one sided on a travel booking side, and it was very canned reporting. We weren't interested in just what Cardinal Health was doing on a global level; we wanted to know what each department was doing."

Buyers also might choose to build a proprietary tool because some of the data they want to manage is sensitive. Some internal data, such as company revenue, might be too sensitive to provide to a third-party supplier, travel consultant Margaret Brady said. And some suppliers also may not be equipped to consolidate data streams. What buyers really want, she said, it to "string them together to find the 'dark data' and make it applicable or smart to your business."

Gisborne also decided that building its own tool would be a more sustainable solution than adapting a supplier's tool. "We could have taken something from [American Express Global Business Travel] and manipulated that," Gisborne said. "As soon as you buy something and start playing around with it, you have to question who supports it and who owns it, so we shied away from that."

Yeah, but Building a Tool Costs a Lot, Right?

With unlimited resources, of course, any company could develop its own data solution, but a realistic analysis of the build-versus-buy conundrum considers both immediate and short-term costs, as well as the internal resources that are readily available to build an internal tool. Naturally, large and/or tech-focused companies like IBM will have more resources. Plus, Gisborne noted, IBM could put its solution on the market.

That's all very well and good for a company that has more than a billion dollars in annual global T&E spend, plus teams of tech experts, but building internal systems is not necessarily out of reach for the midmarket, either. In fact, it still can turn out to be the more economical option.

Some data tools are prohibitively expensive for smaller companies, Brady said. Many focus more broadly than the travel program, and unless a company is already using such tools for other data functions, a full deployment might not make sense for the travel department of smaller company.

Building a travel data analysis tool, though, does not require a large team of tech experts. Often, a few employees could dedicate a few days per week to it, Brady said. Employees with expertise in Excel or business intelligence tools are a good start. According to Microsoft global travel and venue group lead Eric Bailey, "Any 24-year-old techie can figure it out, whether it's Tableau or Power BI, in like a day. Once you have the data, literally anybody can do it self-serve."

Frustrated with "bad platforms that didn't deliver," Steven Mandelbaum, vice president of information services for The Advisory Board Co., said building a travel dashboard internally turned out to be the "cheap use case." It took about two full-time-equivalent employees over a six-month period, he said.

Companies also should look at any data tools in use elsewhere in the company. Cardinal Health, for example, built its travel cube from an analytics tool already in place for accounts payable data. "We took that same solution, which was very analytics driven and wasn't what you would call a business intelligence reporting solution, and put the right data in so you could slice and dice it," Eckhert said. "We were able to get very creative with how we built the data structure, and adding our HR file allowed us to drill down to the employee level, manager level and organizational level."

On the other hand, it takes money to keep a DIY data tool up and running. That's why Allstate opted to have its TMC, Travel and Transport, which this summer spun off Data Visualization Intelligence, manage its data, Allstate director of enterprise travel Duane Goucher said. "Historically, once a project is funded and deployed, funding the update, enhancements and maintenance can be problematic," he said. "A non-proprietary solution also allows feedback and best practices from other users versus being siloed into a unique solution. Letting someone else do what they do best and not reinventing the wheel seemed like the better option."

That's the tack Claire Blades took for Veritas Technologies. As the company was in the midst of a divestiture from former parent Symantec, building a tool wasn't in the cards. That said, she needed a more sophisticated view of her data, as she was about to go into supplier negotiations as a new company with no historic data. When American Express Global Business Travel approached Veritas with its Premier Insights tool, Blades saw the potential for an efficient, off-the-shelf implementation.

"It took eight weeks to get the tool up and running with our data," she said, though she underscored that the tool is only digesting TMC and credit card feeds at the moment, and cannot accommodate the company's purchasing card or business travel accounts. But what the tool does do, it does well, she said. "We get a weekly track of our spend instead of requesting reports. And instead of having this conversation post divestiture, we could establish with our suppliers what our [contract] goals were. It put me in a much stronger position with them and I could show them that we were current with our data."

The Payoff

Whether building or buying tools, travel managers drilling deeper into data have found ways to improve their travel programs. Cardinal Health has brought in TMC and expense data to identify noncompliance, Eckhert said. In particular, the new tool provides visibility into car rental and mileage reimbursement activity, where Cardinal previously had no visibility, and lets the company see when it is reimbursing those who are not compliant, he said.

The increased visibility also has freed up the travel team's time, said Jill Huffman, Cardinal Health senior manager of global travel and meetings management. "If we wanted to talk about a hotel that wants in our program or an airline that wants more segments, we can look into our system and see in real time what is happening," she said. "Or with [traveler] questions, we need to be able to answer those quickly or it's past the point. We knew we wanted to see lost savings, but as we saw the gaps and the things we wanted to see at the end, we started tweaking that [tool] and putting in more and more dimensions and things we wanted to see." The next step will be to make the data interpretation more intuitive by leveraging IBM's Watson technology, she said. "We just gave [IBM] all our travel data, and they're looking at integrating what we did with our travel [tool] and now using Watson as a part of that," she said.

IBM's own travel team also will gain time thanks to the team's internally developed tool, Gisborne said (different from the larger IBM's Watson technology). IBM's travel team previously had a repository that received Excel files but could not report that data in real time or coordinate with supplier data that came in separately. Now, all those manual processes will be automated. IBM's travel data tool also will pull in data, such as on-time flight information for better insight into airline partners' performances, he said.

Blades has configured the Premier Insights tool with Veritas' hierarchy data and is looking at how she can optimize her program. "The drill-down is excellent. I can look at [the behaviors of] my top travelers, my road warriors, and maybe I message them differently or provide something different. It's not all about catching people out but about managing the program better." Blades is also keen to configure the tool with budget goals that will allow businesses to track spend against goals in near-real time.

Although Blades' team has a lot of new intelligence to work with for now, she admits that not having additional data streams coming into the TMC tool eventually will be a drawback. "We've been very transparent with American Express GBT. They have an opportunity to [lead] the industry on single-source analytics. But there is a lot of competition on the procurement side, and they have to start pulling in additional feeds if they are going to compete on a larger scale."

Expanding Opportunities

Once travel buyers bring data sources together to achieve their sourcing goals, Brady said, they can think about repurposing that data for other benefits. A company could, for example, use its tool to look not just at compliance but also at carbon emissions, she said.

Above all, however, buyers need to be sure of their tool's integrity before reporting data or using it to alter policy, Gisborne said. What's most important is "making sure the data is there—and the expense data is key—making sure it's embedded and matching," he said. "If you're going to give info to your CFO that people are traveling in business class when the policy is economy, you want to make sure that is accurate. If not, you've lost all credibility." And those who choose to ignore data completely are missing opportunities. "I invite people managing travel to have intellectual curiosity and understand the path of data," Brady said. "If you don't have an interest in that and just want to do service, you are really missing something."