A few weeks ago, Richard Lee and I caught up with the analytics team behind Rearden Commerce/Deem’s travel solution and platform. On the call, we dug into ways that Rearden is beginning to inject a new set of useful predictive analytics into the business travel area. Their approach incorporates the level of quantitative and correlation analysis of historical data (both personal and publicly available) that makes even a CMU graduate’s head spin (Richard’s b-school alma matter). Among other dataset combinations/mash-ups layered with predictive analytics on top, Rearden is leveraging Bureau of Transportation statistics covering 67 million flights in the past decade and using this information as a foundation to deliver personalized travel analytics to assist with decisions at the point of booking and travel.
The platform is thus able to predict and recommend connecting schedule, flight dates and times, specific carrier and flight number, etc. The approach goes far beyond the “on-time” recommendations for single hops that Orbitz (Richard’s former employer) and other travel sites provide by taking into account the realities of actual travel conditions. It’s somewhat analogous to some of the work SAP has done in the supply risk area by applying predictive analytics based on more than just single-site (and company) supplier performance ratings with the InfoNet offering.
Yet Rearden is embedding this capability within their core offering rather than delivering it as a stand-alone service or solution. And they’re hoping to do this pervasively across the platform, wherever accessing such data is feasible. For hotel stays, Rearden is taking a big data approach that enables the platform to personalize recommendations for travelers based on historical booking/usage pattern, identified preferences, etc. while still leveraging structured information from OTAs like Orbitz, which possess technical details on over 100,000 properties and their holding companies.
The platform that powers this analysis and recommendation is Rearden’s data/scoring engine that leverages self-learning or artificial intelligence (AI) capabilities. It’s part of Deem. With each new piece of data, it adds to the normalization of existing foundation and further narrows down the success rate of predictive forecast. The result is an integrated and intuitive online purchasing experience. Rearden is light years ahead of where Richard was when he was at Orbitz trying to build a fledging B2B business — at the time, with some of the best underlying technology bits available under the sun (or at 35,000 feet).
Flash forward less than a decade and big data is finally hitting B2B corporate travel. Yet the power of predictive analysis far transcends making business travel better for employees. The real bottom-line value is in mining through data and “mashing-up” different datasets to analyze savings, compliance, rebate claims and related information to gain the savings upper hand. And imagine systems doing this for you and presenting opportunities proactively versus relying on sourcing analysts to hunt for things on a periodic basis.
Perhaps Rearden will bring this to users in the future. But believe providers like Opera Solutions and IBM — which each employ hundreds of statistic PhDs and are doing some pretty nifty things behind the scenes in their procurement and supply chain practices — are well on their way to bringing the power of big data to a new savings and compliance level. Still, we hope the individual traveler benefits in the future as well. And for this, we have no doubt that we’ll all owe Rearden a big debt of corporate travel gratitude.
In the meantime, even though we still might be waiting some time for predictive analytics combined with typical P2P environments for frontline users making shopping and buying decisions, we’re already seeing integrated data and contextual recommendations in other areas of procurement suites, including vendor management systems for contingent labor. And we’ve heard of a number of companies (and vendors) doing some interesting things with OSHA and DOT data in the context of spend analysis mash-ups.
Still, putting the power of data into the buying decision itself is a cost management and outcomes-based holy grail that could change the basis of how companies evaluate, deploy and use end-to-end procurement technologies, from upfront spend analysis and strategic sourcing to contract management and transactional purchasing execution. As Ariba, Coupa, SAP, Oracle and many other P2P providers begin to think more about the type of data they could push to frontline users in the shopping and buying process, we’ll begin to see an increased focus on scoring systems, recommendations and metrics based on areas like the personal item-level risk associated with a purchase (e.g. ratings to show a type of notebook, peripheral or mobile device has been returned for service — based on asset management and warranty claims information — more than others that are on an approved list).