Publication Type

Master Thesis

Publication Date



In hotly contested product categories dominated by a few powerful firms, it is quite common for weaker or late entrants to focus only on particular segments of the whole market. The rationale for such strategy is intuitive: to avoid direct confrontation with heavy-weight firms, and to concentrate in segments where these weaker firms have comparative advantages. In marketing, this is what people called “go niche or go home”. The niche-building strategy may rely on “homophily”, which implies that consumers in a particular market segment might possess certain set of attributes that cause them to appreciate certain products better (in other words, weaker firms would customize their products to target some particular market segments and not the mass market). On the other hand, the niche-building strategy may also rely on the network effect, which implies that consumers having social relationship would reinforce each other via their respective adoptions. In this case, weaker firms should recognize such inter-customer network and concentrate only on customers belonging to certain set of strategic clusters. In this thesis, we present the model for building effective niche-seeking strategies as a sequential adversarial search problem (game) on an infinite horizon. For simplicity, we assume that the adoption choice depends only on the network effects (in other words, a customer will choose the product that is chosen by the majority of her neighbor). The social network is directed, and there will be two firms, one with significantly more marketing budget than the other firm. Firms take turns making investment choices on which customer to convert. For both firms, their budgets are fixed over time and unused budget will not carry over to future time periods. Furthermore, we introduced a contractual lock-in constraint to represent an obligatory policy imposed by companies for preventing their customers from defecting to adversary’s product choice. With this model, we manage to show that a simple strategy based on the evaluation of individual customer’s “value” can effectively identify and secure niches within randomly generated scale-free networks. We also show that such niche-building strategy indeed performs better in the long run than a myopic strategy that only cares about immediate market gains. We deployed the Minimax approach to strategically reason the two node selection strategy, and employed a simple α-β pruning mechanism to improve search performance.


consumer choice, customer value, niche strategies, sequential search, influence maximization, influence network

Degree Awarded

MSc in Information Systems


Computer Sciences | Operations Research, Systems Engineering and Industrial Engineering | Strategic Management Policy


LAU, Hoong Chuin; CHENG, Shih-Fen