Challenge
AMS Group is a cohesive group of established companies that provide technology and security equipment to aerospace, defense, and security markets. AMS discovered that one of the companies within the group was sitting on an untapped data bank of procurement RFIs, RFPs, and awards going back decades. The data had the potential to reveal actionable insights into the trends and patterns of government buying behavior, from the nuts and bolts needed for warship maintenance to multi-billion dollar military helicopter procurement projects.
AMS Group had a potential goldmine of information at its fingertips, but lacked the technical and analytical capabilities to derive any insights from it, so they partnered with EastBanc Technologies to discover what they might be able to do with the data.
Solution
Faced with so much data, we began by time-boxing the project to help focus on value (time-boxing is an iterative approach that emphasizes the incremental delivery of a solution, from the start of the project, instead of trying to deliver it all at once and is a key tenet of Agile software development). Next, a Minimal Viable Prediction (MVP) was established. An MVP is an outcome that addresses a primary problem and is the starting point for any journey to predictive analytics - whether it’s predicting defense purchasing needs, future revenues, or the lifecycle of military equipment, whatever it may be, we focused on finding ways to get to that in the fastest way possible.
We discovered a breadcrumb trail of data that could shine a light on the key elements that comprised a winning bid, such as the combination of a material, price, time of procurement, and perhaps a certain middleman, etc. When combined, a pattern emerged. It became clear that these factors were consistently present in winning bids, while other combinations resulted in a lost opportunity.
Leveraging this data, testing this assumption and working towards an initial MVP of “which factors determine a winning bid?”, EastBanc Technologies developed an initial version of a predictive analytics engine for AMS Group. The business impact is significant. Not only does the engine save participants in AMS Group potentially millions of dollars in chasing low probability bids, but also provides a trickle-down revenue stream for AMS Group through onward sales of the insights the engine can derive to third parties, such as other government contractors.
Once the MVP is achieved and this nucleus is in place, AMS Group can build from there. Iteratively tweaking their data, adding new data, or incorporating new technology.
Business Impact
To date, the solution has been in place for 12 months and is generating revenue values in the millions of dollars per year. The company’s first quarter conversion rate for quotes to orders has improved from an average of 5.2% to 12.2% year-over-year, while the average order value has also increased by a stunning 94% ($8,268 to $16,112).
Technologies Used
R Studio, Azure ML Studio.