Recommendation-based retail is an emerging business model in which online retailers proactively ship products to consumers based on predicted preferences, allowing consumers to purchase only items they like. Unlike traditional retail, where consumers actively search for products, recommendation-based retail reduces consumers' search efforts but carries the risk of inaccurate recommendations. Despite growing adoption in practice, strategic implications of retailer competition on adoption decisions remain unexplored. Using a stylized game-theoretic model, we analyze competition between two online retailers considering whether to adopt recommendation-based retail. Equilibrium analyses identify how adoption costs (infrastructure investment) and recommendation accuracy influence retailers' strategic decisions. We present three key findings: First, even under symmetric conditions, an asymmetric equilibrium emerges where one retailer adopts the recommendation-based model and the other does not. Second, retailers can face a prisoner's dilemma, choosing a Pareto-inferior equilibrium despite mutually beneficial alternatives. Third, consumer welfare improves as recommendation accuracy surpasses a certain threshold.