Customer engagement is a bit of a game, because, deep down, it’s a form of haggling and bargaining. Let’s be blunt: everybody has an ulterior purpose and is manipulating the other party in that direction. The customer is trying to get the best deal from you, and you’re trying to hold onto them and sell them more stuff at a healthy profit.
Customer engagement is not solitaire, and, unlike many online games, it always has very real stakes. By its very nature, customer engagement is an interactive decision process involving individuals and organizations, entailing varying degrees of cooperation and conflict in the course (hopefully) of a stable and mutually beneficial outcome.
Game theory is a modeling discipline that focuses on strategic decision-making scenarios. It leverages a substantial body of applied mathematics and has been used successfully in many disciplines, including economics, politics, management and biology. There has even been some recent discussion of its possible application in modeling customer-engagement scenarios to improve loyalty, upsell and the like.
Customer engagement modeling is a largely unexplored frontier for game theory. The literature on this is relatively sparse right now, compared to other domains where game theory’s principles have been applied. To jumpstart the discussion among data scientists, I’ve prepared a few thoughts for how they might apply game theory to engagement modeling. Note that I’ve substituted the word “engagement” in place of “game” in describing several key approaches. The core concepts translate over to this application domain quite well:
- Cooperative engagements: These are where interaction is in the context of an established relationship. Customer retention, upsell and cross-sell are examples of these sorts of engagement. In a cooperative engagement, there is the possibility of substantial communication between the customer and the business in question as they come to terms. The conversational give-and-take of B2C next-best-offer communications are a good example of cooperative engagement. In a non-cooperative engagement, however, there is no established relationship and opportunities for information exchange between parties are minimal. Marketing-response interactions with social-sourced customer leads are an example of non-cooperative engagements.
- Variable-sum engagements: These are where the summation of the participants’ respective outcomes–gain and/or loss–can range widely, depending on how good or bad a deal it is for them. By contrast, constant-sum engagements are where the sum of gains is some fixed amount, usually some positive number or, worst case, zero. Positive constant-sum engagements are an interesting possibility for marketing and promotions, in the form of jackpots, lotteries and other bonanzas that are divvied up among one or more lucky winners. Customer engagements are rarely zero-sum, because the customer (hopefully) gains a value greater than what they spend and the business (hopefully) profits as well.
- Asymmetric engagements: These are where the returns to the various participants depend on their identities, not on the specific strategies they’re pursuing within the relationship. One common example of an asymmetric engagement is any B2C relationship where the customer identity is known and in which offers are targeted precisely to that customer’s profile, history and propensities. Symmetric engagements, by contrast, produce the same outcomes regardless of the identities of the participants. Sealed-bid auctions, more often used in B2B than in B2C, are a good example of such an engagement strategy.
- Sequential engagements: These are where participants have some knowledge, not necessarily complete or perfect, of earlier actions taken by others. This is a typical CRM assumption, where both the customer and the business take actions in the context of an ongoing relationship that involves a never-ending string of interactions. The assumption may also kick in where no such relationship exists. For example, it applies in cases where market research gives the business some insight into the prospective consumer and the customer has some knowledge of the business’ track record and reputation on engagements similar to the one they’re contemplating entering. Simultaneous engagements, by contrast, are where all participants act at the same time without knowledge of the other’s past actions or track record. This latter assumption may apply to engagements that involve one-off spot transactions between strangers where a third-party vouches for all participants and guarantees transactions.
- Many-player engagements: These are where there is an arbitrary, but finite, number of customer participants who consider each other’s actions when deciding how they themselves should act. This assumption applies to any B2C engagement that involves making differentiated offers contingent on how the consumer’s “friends and family” have bought. It is also relevant in any online engagement where the recommendations or activities of “influencers” (friends, family, experts, Oprah, etc.) have some bearing on the choices we make. And it describes, to some degree, the business model of social-shopping communities such as Groupon. By contrast, single-player engagement (i.e., where there is just one customer per engagement) is the predominant CRM assumption for most customer-facing scenarios. Most B2C “gamification” experiences are in fact geared for single-player engagements; they provide fun interfaces and bestow “rewards” to encourage individuals to engage in otherwise unrewarding tasks, such as completing surveys, completing forms and looking at display ads.
At first glance, data scientists may consider themselves fish out of water when it comes to applying game-theoretic approaches to customer engagement. Methodologically, game theory looks at discrete variables–actions, events and outcomes–rather than the continuous variables that are the heart and soul of data science’s core discipline of regression modeling. In addition, game theory assumes that we should model engagements as interactions among rational decision makers–individuals and businesses–that can have deterministic outcomes, rather than the probabilistic outcomes associated with mainstream data science.
Down deep, game theory is the realm of what some have called “decision science,” rather than data science in its traditional sense. Nevertheless, it provides a valuable set of approaches for behavioral analytics. Game theory can deliver rich insights, especially when data scientists use it to enrich and extend the propensity, experience, and other behavioral models at the heart of customer engagement.