This article proposes a method to uncover opportunities for exploitation and exploration from consumer IoT interaction data. We develop a unique decomposition of cosine similarity that quantifies exploitation through functional similarity of interactions, exploration through cross-capacity similarity of counterfactual interactions, and differentiation of the two opportunities through within-similarity. We propose a topological data analysis method that incorporates these components of similarity and provides for their visualization. Functionally similar automations reveal exploitation opportunities for substitutes-in-use or complements-in-use, while exploration opportunities extend functionality into new use cases. This data-driven approach provides marketers with a powerful capability to discover possibilities for refining existing automation features while exploring new innovations. More generally, our approach can aid marketing efforts to balance these strategic opportunities in high technology contexts.

翻译：本文提出了一种从消费者物联网交互数据中挖掘利用与探索机会的方法。我们开发了一种独特的余弦相似度分解方法，通过交互的功能相似性量化利用机会，通过反事实交互的跨能力相似性量化探索机会，并通过内部相似性区分这两种机会。我们提出了一种拓扑数据分析方法，该方法整合了这些相似性成分并实现其可视化。功能相似的自动化揭示了替代用途或互补用途的利用机会，而探索机会则将功能扩展到新的应用场景。这种数据驱动方法为营销人员提供了强大的能力，既能发现优化现有自动化功能的可能，也能探索新创新。更广泛而言，我们的方法有助于在高科技情境中平衡这些战略机会的营销工作。