In sales environments with a direct selling force, territory management and quota setting is an essential process but is often done manually with limited real insights behind the qualitative assessments. Data availability in this context has been devoted to enhance this process logically, explicitly, and transparently. In this spirit, following the first offer of a measurable workflow (i.e., the Generalized Flow), predictive techniques in modeling the winning propensity are explored to support the qualification of the depths of the sales funnel as well as the quota setting process in salesforce allocation, evaluating the theoretical and practical value of its implementation. In a salesforce quota and territory optimization flow, three fundamental parts are addressed with industry cases: the lead grouping much widened by data-driven building rules and technique integrations, the territory management with optimized routing paths both in the existing lead distribution process and time added contacts, and the allocation of quotas wrapping each location and salesmembers. This sales prediction workflow stands at the meeting point of data science, sales action pipelines, and business understanding.

Along the predictive move, the first step involves knowing the background of the sales industry and the predictive process, as well as the opportunity construction and sales technology in the business; the second step is to fully leverage the acquired information of that business environment and the insights of the data; the third step is to prepare and preprocess the features as predictors for the winning propensity prediction. Once the geographical segmentation is decided, the attention will move to assign sales personnel to each territory in the sales quota settings. The territory management problem is defined as MCSPP with minimum contacts to cover maximum potentials along the rerouted length constraints, and solved by a heuristic algorithm based on heuristics and a non-linear programming solver. Implementing this systematic approach will greatly alleviate the burden of quota settings, such as blind manual adjustment of quotas over territories or inaccurately fixed quota ratios across salespersons. Instead of being arbitrary or with limited quantitative support, quota distribution can be conducted on a quantitative basis. The territory management and quota settings in this study are tailored for business adoption by fitting the data environment of the state-of-the-art analytics