What Game Theory Can Teach Us About Media Budgets, Agency Reshuffles And Attribution

“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Tom Triscari, co-founder and managing partner at Labmatik.

Every October, the Nobel committee hands out its annual prizes in recognition of academic, cultural or scientific advances. While there is no award for advertising, that doesn’t mean Nobel laureates don’t offer valuable lessons for advertisers.

Two recent Nobel Prize winners in economics – Richard Thaler and Lloyd Shapley – are particularly relevant to three of the biggest advertising challenges: managing limited budgets, partner alignment and transparent partner compensation.

Thaler, from the University of Chicago, won this year’s Nobel for revolutionizing economics by upending assumptions of humans as rational, self-interested actors, instead showing how human psychology and irrationality impact economic decision-making. To me, Thaler’s assumption of humans as predictably irrational is something we can certainly apply to the current state of media budgeting and advertising in general.

One of Thaler’s most famous concepts is mental accounting. Thaler “showed that people do not regard all money as created equal,” The New York Times reported. “For example, when gas prices decline, standard economic theory predicts that people will use the savings for whatever they need most. In reality, people still spend as much of their money on gas by buying premium gas even if it is bad for their car. In other words: They treat a certain slice of their budget as gas money.”

Household expenses like gasoline are very similar to media budgets. While CMOs rightly rail against uncontrolled fraud, unmeasured impressions, inaccurate viewability metrics and non-transparency, digital and programmatic ad spend keep growing like mad.

The same irrational dichotomy can be found in the TV world. Viewership continues to decline, making it harder to reach target audiences, yet marketers nonetheless sustain TV budgets, paying increasingly inflated rates for TV ads, even as CPMs fall in other channels and reaching audiences on TV becomes increasingly difficult.

Clearly, mental accounting has infiltrated corporate ad budgeting and its sell-side counterparts, flaws and all. While rational ad buyers should embrace channel-agnostic buying models, avoid upfront commitments in favor of greater budget flexibility and make real-time impression pricing decisions based on predictive variables, real humans favor the irrational comfort of channel-based budgets, quasi-performance guarantees and upfront-negotiated, fixed pricing. Thaler would have predicted as much.

How Market Design Predicts Agency RFPs And Budget Cuts

Thaler isn’t the first Nobel laureate to create concepts relevant to advertisers. Lloyd Shapley, along with Alvin Roth, won the 2012 Nobel prize in economics and revolutionized how economists think about market design, matching theory and payoffs. (Disclaimer: I took Shapley’s class as an undergrad.)

Shapley’s matching theory states that if rational people – Thaler’s mythical creatures who always know their best interests and behave accordingly – simply engage in unrestricted mutual trade, the outcome should be efficient. If it is not, some individuals would devise new trades that make them better off.

You can see this trade-off playing out today in the client-agency relationship. On the one hand, agencies and other managed service partners (aka the “ad stack”) claim to provide substantial value-add, yet they are unable or unwilling to disclose the real price for these unmeasured services. When the advertiser does not know the true cost, it is unable to accurately price the purported value-add and determine if the deal is good or bad.

In response, big clients are cutting spend and reducing the number of partners on the plan. They are also contracting directly with ad tech vendors and bringing more processes under their control. In essence, some (rational) advertisers are doing what Shapley would predict by devising new trades.

Shapley Provides The Solution

Another Shapley idea is known as Shapley values — a concept for proportionally dividing the benefits of cooperation among multiple participants according to individual contribution. The current state of the programmatic puzzle could not be better described.  

For example, imagine a coalition of players, such as marketers, creative agency, media agency, ad tech, data providers and publishers, cooperating to create value from that cooperation. Some players may contribute more to the coalition than others, and some may possess advantageous bargaining power, such as kickbacks, rebates or undisclosed arbitrage. As such, Shapley would ask, “How important is each player to the overall cooperation, and what payoff can each player reasonably expect?”

The key ingredient to solve this puzzle would be to invent a common method to calculate value creation. This is, of course, easier said than done, but what if you found out that great creative trumps all the other contributions? What if bidding on users instead of inventory drove the remaining value-add? If so, all the other players might not matter so much.

AI Reduces The Sprawling Advertising Apparatus

The rapid development of artificial intelligence like Albert AI and IBM Watson has more than a fair chance to improve irrational advertising sooner than we all think, and programmatic sits at the center of it all.

The idea is simple: Plug in your budget, set a well-defined singular objective and the machine will make rational decisions to achieve it. Moreover, connect the machine to blockchain, and a rational ledger is generated to account for all decisions showing who and what contributed to the gains or losses.

Don’t be fooled – AI advertising will look more like a silver bullet if humans wire the machine to mirror what they already do because in a perfectly rational cognitive computing world, marketers will still need to ask: Does the human know what to do? Can the human get the machine to do what he or she knows?

Follow Tom Triscari (@triscari) and AdExchanger (@adexchanger) on Twitter.

This post was syndicated from Ad Exchanger.