Analytics for the Engagement Life Cycle of IBM’s Highly Valued IT Service Contracts


IBM competes to win multi-million IT service contracts. These large contracts typically involve composite services composed of several thousands of software, hardware, and services. Examples are data center consolidation, migration of IT services to the cloud, and help desk services management. In response to clients’ request for proposals (RFPs), IBM and other competing IT service providers submit proposals. Clients short list a number of providers and engage with them through intense negotiations to select a final winner for the bid. Service providers maintain and manage a pipeline of such deals. Each deal goes through a life cycle which begins with the identification and validation of the opportunity, qualifying it, receiving a RFP, pursuing it with a team of business and technical sellers until the contract is signed or the deal is not won.

Given the business value at stake, the conventional approach to taking these steps involve resource-intensive, and complex activities and decision making. This calls for a strong demand to bring in data-driven analytics to help manage the pipeline, make resource allocation decisions, strategize the winning of each deal, and forecast the revenue out of contract signings.

The team at IBM Almaden Research Center has partnered with stakeholders in IBM services organization and developed an analytical toolset that offers insights during various stages of this life cycle to assist different decisions. In this presentation, we will describe the objective and overview of the following tools, their key innovations, and resulting business impact:

• A requirement analysis tool that uses innovative NLP techniques and data mining to analyze clients’ RFP documents, extract requirements, and map them to relevant offerings.
• A pricing tool that consists of data mining of historical deals and market data in order to calculate pricing points, and a predictive model that provides the relative win probability of each price point.
• A work in progress resource allocation tool that optimizes the assignment of the sales personnel to the negotiation of different deals.
• A deal competitiveness assessment tool for what-if scenario studies, price re-assessment, and analyses of “in-flight” deals w.r.t. similar won historical deals.
• A win prediction tool that combines a quantitative predictive model and a text-based one that analyzes the comments written by the sales team during the pursuit.
• A predictive tool for the type and timing of the next milestone to be achieved in the engagement life cycle.
• An aggregated revenue prediction tool based on optimizing the weights on aggregated historical quarter revenues of deals at different stages.

As can be seen from the description, the tools span over the different types of analytics (descriptive, predictive, and prescriptive) as well as combinations across methods. The tools are innovative in their overall framework, detailed algorithmic design, or both. The developed tools have already been implemented, deployed, and are actively being used by the business. The impact of using these tools is significant to the business. The expected revenue gains are dramatic, projecting a multi-million dollar increase in revenues and more efficient management of the overall process.