Commentary

The "If & Then" Algorithm In Marketing

The most dramatic advances in human development over the past century can be summed up in two words: “if” and “then.”  With only slight oversimplification, all marketing can be translated into if/then processes. 

Look no further than Moore’s law or IBM’s Watson. The algorithm, or the ability to perform complex tasks with a cascade of simpler if/then actions, has yielded stunning advances in machinery, medicine, communications, and artificial intelligence that are redefining our lives on several levels. 

It may seem a stretch to tie those lofty heights to the earthier world of marketing. Yet it’s as true for marketing as it is for any human endeavor that future advances lie in more fully harnessing the force of if/then. 

advertisement

advertisement

The If/Then in Programmatic Media 

We see the early effects of if/then in the rise of programmatic media. Programmatic advertising allows marketers to perform instantaneous if/then decisions across millions of potential customers and thousands of destinations. It is now a question of when, not if, all media will be programmatic. As significant as that is, media purchase and delivery is not the full extent of programmatic’s impact. 

In fact, if/then harnesses the two areas that are currently most at the tip of every CMO’s tongue: data and content. That’s because data feeds the “if,” and content provides the “then.” Data signals the people and the context around them that spur an adept marketer to act. Content is the action the marketer takes to react to that signal. 

Algorithm ExampleIf The Home Depot sees a suburban homeowner within a mile of their store on a warm April day, then it forwards spring planting tips and an offer from their gardening center. The response generates data that sets up the next if/then. If the spring planting tips led to a purchase, then gauge their interest in a loyalty program for the summer ahead. The effectiveness of the marketing correlates with the depth and breadth of the if/then branches the marketer can meaningfully define. 

The Art vs. Science If/Then Battle

The most successful brands are those who inspire deep-seated passion and loyalty like Harley-Davidson and Nike. They exist in an emotional territory above the mechanical abilities of if/then algorithms. The Art vs. Science group might argue that applying the if/then paradigm to its logical extreme overlooks the human factor that underlies the best marketing. Of course, they’d be wrong. 

Algorithm Example: If/then unleashes creativity rather than constrains it. It’s true that you’ll get mechanical transactional marketing if all you feed into the “if” is mechanical transactional data. But if you feed individual interests and passions into the “ifs,” you uncover ways to forge even more meaningful human connections. Imagine how much more value a marketer could deliver to a runner who just had a child, or who’s on a business trip away from their regular running route, than it could to just a runner who is due for a new pair of shoes. 

The if/then algorithm approach allows marketers to deliver at a scale and sophistication that evades today’s most intimate marketer.

This topic was inspired by a discussion with DigitasLBi’s Chief Strategy and Media Officer Baba Shetty, a brilliant thinker on various topics, including commercial media.

1 comment about "The "If & Then" Algorithm In Marketing".
Check to receive email when comments are posted.
  1. George John from Rocket Fuel Inc, June 10, 2015 at 11:45 a.m.

    My experience is that simple if/then rules designed by people can feel very empowering at first, but then people start trying to optimize for better results and it becomes an un-maintainable mess.  I think an ideal marketing plan will have marketers defining specific rules in rare cases for triggered or situational messaging (for example, the starbucks app on my phone asks me if I'd like to tip the barista shortly after paying), but for more mainstream performance-oriented messaging there will be some constraints set by marketers (for example, certain offers are only eligible for gold customers) and then an AI system will be allowed to show whatever message is most likely relevant to the customer and ROI generating for the enterprise.   You can think of that AI system as learning its own rules based on what works empirically, but they're more complex and maybe surprisingly allow for more shades of grey and nuance than human rules.  In our history of media buying at Rocket Fuel with nearly a billion in total revenue to date, I only remember one time in 2009 when a human if/then rule beat the machine-learned scoring models.  It was for a free internet voice communications app, and we experimented targeting people in the US whose browser language was not set to English, realizing that people would likely use it to communicate internationally.   

Next story loading loading..