• by May 6, 2005

By Joseph Zawadzki

Motorcycle maintenance gets Zen. Business gets Sun Tzu. We in marketing will have to settle for poker. In looking at America's favorite pastime, there are some valuable lessons that can be applied to behavioral targeting and online advertising.

Make Good Decisions Based on Bad Information

Mike Caro, the Mad Genius of Poker, starts his seminars with a question designed to put students in the right frame of mind: "What's the goal of poker?" When someone obligingly responds that it's about winning money, Caro has them right where he wants them. Poker is not about winning money, taking "pots," or even outwitting your opponents. Poker is all about "making good decisions." That is the only thing you control in the game of poker, and - over time - is the one that allows skill to overtake luck.

You can't determine how the cards are going to come out or how someone else is going to bet, but based on the available information - what cards you have, what you think your opponent has, the size of the pot, how many of a certain card are left in the deck, etc. - you can make the best possible decision despite imperfect information.

Sound like online marketing to anyone else? Online data can be pretty spotty. There are hundreds of sources that vary greatly in terms of coverage and accuracy, forcing marketers to make the best decision possible without a complete or perfect data set. The trick is to take people you do know a lot about - typically your customers - and make general assumptions about your prospects using anonymous data. If you do a good job with the "birds of a feather flock together" approach, and start treating the anonymous visitors as if you knew a lot about them, then you can realize some real gains.

Learn to Recognize Consumer "Tells" Poker "tells" are those unconscious mannerisms that provide an extra insight into the quality of a player's hand. Some people are actors and project strength when they have a weak hand, or project weakness when they are in a commanding position. I play with one guy whose hands shake when he bluffs and another whose hands shake only when he has an unbeatable hand. Understanding what that same action means for each of them is what makes the "tell" usable.

In online marketing, you begin to establish a history for individuals as you delve into their behavior on landing pages, shopping carts, member areas, and more. This unique data set can be used to learn key patterns - or "tells" - that drive response. The challenge, like in poker, is drawing the right inference from the available data.

Collecting the right consumer data and then understanding what it means for you specifically is the best way to predict future behavior. Based on that understanding, you can then begin to make good decisions and improve your overall results.

Work Backwards to Become Better Marketers The trick with poker is to work backwards. When someone turns over their cards at the end of the hand, good players try to recreate how the hand played out - putting it into the larger context of how everyone else played the hand and how that player handled similar situations in the past. From that, a model for future behavior can be predicted and you are in a better position to take advantage of the situation. A similar methodology can be applied to behavioral marketing. Rather than simply buying a segment and hoping that it works for you - work backwards. Look closely at your responders to understand which behaviors they exhibit and then go out and buy those.

Keep the Game Moving Unfortunately, neither poker players nor marketers have forever to make decisions. So when applying the lessons above to your online marketing campaign, remember not to succumb to analysis paralysis. Just like some poker players will endlessly debate the merits of a raise versus a call with a pair of Aces in early position, many marketers will spend too much time debating whether banner burn-out is taking place on a site that represents only 2 percent of the spend.

Since that pair of Aces in early position happens only once every 1,100 times, and that the theoretically "correct" answer to both this and the burn-out question might only increase aggregate performance by a fraction of a percentage point, the important thing is simply to make a quick decision and move on. Balance speed and accuracy by focusing on the right projects with the most potential at the lowest cost, and even though you won't be right all of the time, over the long run it will prove to be your most profitable strategy.

Joseph Zawadzki is founder and president of Poindexter Systems.