It seems that some corporate travel decision-makers are endlessly dissatisfied with data. It’s imperfect, dumped on them in buckets and hard to present to executives. Nowadays they want insights at the trip level. They want to identify policy violations before it’s too late. They want to take more immediate action on predicted outcomes.
Using travel data for predictive analysis and benchmarking is this year’s top priority among 1,113 travel, meetings and events professionals surveyed last fall by Carlson Wagonlit Travel.
In some cases, travel departments are looking to business intelligence firms that came of age outside the travel industry and now are targeting it. PI Ltd of London is an example. After working for three years with BCD Travel in Europe, the company last month hired former TRX exec Tom Tulloch to expand in North America. Like others including Mo-Data, PI started in other domains before corporate clients brought it to travel data.
Elsewhere outside the travel industry, tools from Domo, Qlik, Lavastorm, Pentaho and Tableau offer self-service opportunities for companies to up their travel data games.
“Those work if you have solved the underlying data problem,” said Brian Beard, Travel and Transport’s general manager for strategic initiatives.
Industry data management specialists have been working on that for many years.
Concur had a couple initiatives going with the former TRX TravelTrax and its own Cognos-powered Insights. SAP owns Business Objects, like Cognos a business-intelligence stalwart. A press official declined to comment on Concur’s plans, but expert sources outside the company thought it likely that SAP/Concur would build a new product in the next couple years.
Grasp Technologies claims it’s the biggest consolidator of T&E data, with more than 600 clients. Grasp established itself by connecting its software directly to TMC back offices, improving data quality. Now, a big part of the firm’s business is consolidating from multiple sources including card, expense, booking, GDS, trip approvals, HR and ERPs. “A while ago, people just wanted total spend,” said vice president Dave Lukas. “Then they wanted it by air, car and hotel. Then by airline. Then by city pair. Then by person. I still think we’re at the infancy of data consolidation as far as doing it right and making it actionable. We’re still finding ways to make it better after 15 years.”
Founded by former TRX exec and Hi-Mark Software founder Kevin Austin, Big Data Experts expects to roll out of its garage in about three months. Hotel commission specialist CTS Systems and dining network Dinova already are using certain components. The company recently hired industry veteran Todd Kaiser to ramp up its marketing efforts.
What’s Wrong With The Data We Have?
It’s a common question from senior execs asked to sign off on a budget for data cleansing and consolidation. They may believe that what’s available from travel management companies is good enough.
Many TMCs invest in data reporting. CWT, FCm Travel Solutions and Travel and Transport are just a few that now are developing new data solutions. CWT’s is due out this year and Travel and Transport hopes to demo a prototype in three to five months. American Express Global Business Travel now is rebuilding much of its technology and points to its relationship with its former parent as an advantage in accessing payment data alongside travel.
TMCs also are frequently criticized for a lack of technological savvy. One former TMC official who declined to be identified doubts their capabilities. “We poured a lot of money into it, but it was rare that we thought we got something right,” said the source. “I don’t know that anyone is better than anyone else among the big TMCs.”
Corporate travel data challenges are well-documented. Accuracy is a huge problem. Timeliness is another issue as some of the most commonly used data sources typically produce information only every few weeks. A lack of standards means some data are input manually, introducing human error.
These issues compound tremendously when corporate hierarchies are involved, or at multinational firms or those using multiple TMCs. It’s harder, too, when multiple data sources are required.
“Most companies continue to rely on one source of data to measure what’s going on with their T&E spend,” said PI’s Tulloch. “That’s mostly TMC data, but there are very valuable reasons why you need to supplement that with card and expense data. Other sources of data are gaining popularity as well, especially social media data, weather, reviews and traveler satisfaction data.”
Further complicating matters, “pretty much everyone thinks they own the data,” said Oversight Systems CEO Patrick Taylor. This creates lag in data acquisition. Data privacy regulations also make it difficult to authorize third parties to receive data. Oversight helps companies inspect travel, expense and payment data for fraud and policy violations.
Taylor said many companies have no consistent “data key” that firms like his can use to inexpensively pull together the relevant data. An example would be an employee ID of some sort. Grasp, too, stresses the importance of keys, while PI cautions against relying on them too much.
“What you can count on,” Taylor summarized, “is that it won’t work perfectly every time. You can be thoughtful up front and build a data model that is robust enough to handle all the different circumstances, and that will help. Other lessons come from trial and error.”
Austin said the challenge defines the necessary data, and it’s important to recognize inaccuracies even if they can’t be corrected. “A lot of what frustrates people is they may never realize how bad the data was,” he said. “They just end up with a bad result.” Austin’s company is building a data scoring module in which the user can weight what’s most important and understand how relevant to a given task certain results are, despite flaws in the data.
Depending on the purpose, there may be nothing wrong with your data.
The way BCD Travel vice president for innovation and intelligence Torsten Kriedt sees it, some clients seek only a full picture of their aggregated spend. For them, a basic level of data reporting will do. At the next level of sophistication, he said clients need “booking and spend data matched on an aggregate level, for example if they want to negotiate with hotels based on total room nights and spend.” At the highest end, clients looking to “understand the full pattern at the trip level or uncover hidden spend, for example around ground transportation or dining,” need something like PI or Concur’s TravelTrax, Kriedt said.
“Clean data and what I can do with it to move my program forward is what I am interested in,” Cisco Systems director of global procurement services Susan Lichtenstein typed during an online conference hosted last fall by the Association of Corporate Travel Executives. “Policy compliance is my largest concern. We use TRX and our TMC for this work. I think the TMC has to get better at presenting data in an actionable format. We always receive data dumps that tell me what I did. I want data that tells me where to go next and why. I want trending information, actionable solutions and projections that will enable a better experience. Also, savings … where am I missing the mark and why and how to solve for it? I want laser-focused information, not a data dump so me and my team have to sift through it.”
Several travel managers in subsequent conversations concurred.
This is the direction Grasp is taking. The company expects to deliver next year on benefits from artificial intelligence and machine learning. Grasp already is helping clients catch issues more quickly. One customer, for example, gets an alert every time an employee submits an expense for laundry on trips of three days or less. “The future for us is a platform showing you opportunities you never knew were there,” said Lukas. “We call it the virtual travel manager.”
Was That ‘Unique Device Identifier’ Or ‘User Defined Interface Data’?
One challenge in picking a non-industry specialist is the time it needs to figure out the jargon and quirks of the industry. It’s debatable how hard that stuff is to learn.
“What I have found with these generalists trying to come in is that they don’t realize how screwed up travel data is,” said Travel Tech Consulting’s Norm Rose. “For instance, you may have five or six names for the same hotel from different sources.”
PI can attest to that. “We thought it would take us six to eight weeks but it took a year or so to really come to grips with travel data,” said PI Ltd CEO Keesup Choe. “It’s a different animal, full of complexity. We’re very interested in this space because the data is fragmented and not orderly.”
Kriedt said requirements for domain expertise lead naturally to partnerships.
Cornerstone Information Systems has worked with Qlik Technologies for years to boost its iBank business intelligence software. Cornerstone vice president for product innovation and marketing Rock Blanco is no longer convinced that industry domain expertise matters much. “It’s really a self-service environment now,” he said. “In the new Qlik Sense tool, you can import a spreadsheet and build visualizations. It also helps with the consolidation side, meshing together disparate data sources like expense, card and back office.” Blanco said self-service is “not for the faint of heart” but many firms now employ data scientists. They may not know lingo but can “look at the data and find new relationships.”
A data scientist at security technology firm McAfee created travel reporting with Qlik’s Qlikview using corporate card, Oracle and SAP data. The company now tracks “who is not complying with a company ‘seven days in advance’ policy,” according to a case study published by Qlik.
That’ll Be How Much?
Self-service of course saves money compared to full service, but a staff data scientist may not be available for travel procurement purposes.
“I haven’t found a data aggregator yet that seems to be able to take multiple data feeds from different sources (pre-trip agency, expense, general ledger, card) and normalize the data globally or provide predictive data and executive scorecards that we can take action on at a development price that is reasonable,” said CDK Global director for global travel and event services Madia Sargent.
Sources suggested the cost of a high-end data consolidator for a travel department can rival the cost of TMC services.
“I’d love to use a company like PI,” said Camilla Lagesen, travel manager at Kongsberg in Norway. “The problem is selling it into my company. I know the data we have are not correct, but the company is satisfied with what we get. Maybe it is good enough, but I believe it can be even better.”
BCD Travel’s Kriedt recognized the concerns. He said clients often hope travel can get in on larger corporate business intelligence initiatives, but it tends to remain “at the bottom of the request bin.” Choe said PI doesn’t simply put people on the problem, because that would be expensive. The company takes advantage of the infrastructure and algorithms it built for the healthcare and retail sectors, which he said handle far larger data volumes.
More affordable solutions, Kriedt said, are elusive for small and medium enterprises because cost savings require standards and scale.
“Pricing varies wildly from client to client,” said Grasp president and CEO Erik Mueller. “We can implement a travel agency with 100 accounts in less than 45 days. When we work with a corporation, just the contracting process alone takes months. Then we have to get through IT hurdles. It’s upwards of 10 to 20 hours just to get approval to connect to their environment and work out data acquisition. We’re a flat fee; we don’t charge by transaction except if someone wants a direct GDS feed. For a corporation where maybe you’re consolidating 50 travel agencies, HR, three credit cards, three expense systems … transaction pricing kills them.”
Additional info: Sources also included five more corporate travel managers, several executives at data management firms and two corporate travel consultants.