AGR has utilized its Big Data advantage and “Analytics Plus” to perform on a series of technical services logistics programs. Our team reviewed Federal government logistics opportunities and concluded, with technology leverage and procurement expertise, a viable project could yield profitable results starting from a ground zero position.
This project reviews an average of over 2000 procurement requirements, 365 days a year and categorizes them first by major sourcing type and then by technical condition. A secondary filter is used to review the procuring agency and to apply a success factor, sifting higher success opportunities for asset utilizationpursuit. These analyzers create a base data block that is warehoused in a master pursuit category.
We then use reverse procurement commodity logic to create a classification model for need to action by logistics professionals. The field procurement execution teams utilize classification models by evaluating their assigned actions into an analytic tool that evaluates against Null, Bayes, and single-variable models to prioritize work packages including supplier interactions. This allows us to maximize our interrogative and execution actions against limited capacity. Since most simple engineering and buy-sell activities at the federal government level produce razor thin margins, adding capacity to process transactions may not be profit effective.
The priority sequence for our classification of action model is:
Phase III of the Logistics Project will be to construct a more rigid “trained” classification model to streamline the prioritization process and reduce model classification analysis cycle time. This classification modeling approach is a classic case of adapting aggressive commercial business practice to FAR controlled government procurement and sustainment. We are also working on the specifications for LOG likelihood module to pre determine success likelihood using a plot of events with the Y axis being True Positive Results and the X axis being False Positive Results that further refine the priority of RFP responses.
Our team has also moved into validation of our models so that we can continue to expand our Federal Government footprint on new data and/or model validation. Our efforts here are first to avoid bias or the systemic errors in our models such as under predicting or over fitting in our model efforts. The data visualization of our efforts are driven by straight forward pragmatic metrics such as revenue growth of this business line versus actual revenue growth, profit margin on an ABC cost basis versus actual resource costs expended, and absolute profit versus target profit.
We have been able to utilize our technical skill base in massive data manipulation without “hands on ” skills as logisticians and procurement professionals to create a high growth consultative business line winning prime logistics and sustainment contracts with the US government.