A Learning Based Vertical Integration Decision Model

Abstract

Vertical integration decisions, also referred to as make or buy decisions, affect firm activity areas that determine boundaries that play a strategic role in the success of firms. The decisions may result in form of spot market usage, long-term contracts, joint ventures, and in-house productions. This study first reviews the decision modeling approaches suggested in the literature for vertical integration. The review shows that for decision making, most of the approaches are qualitative in nature and but there are also optimization models and expert based models. Thus, a new learning based probabilistic classification approach is proposed to model the vertical integration decision problem. This data-driven quantitative approach using Naive Bayes is able to represent the uncertainty inherent in the decision problem. The applicability and effectiveness of the developed learning based vertical integration decision model is demonstrated by an example that covers real software make or buy cases in a retail company. 85% of the decisions were correctly classified by the developed data-driven model.