By Published on .

Optimization is a sunny idea that quickly clouds over when we discover the havoc it will wreak on traditional media planning. Victims of the whirlwind will include CPM buying, daypart selling, guarantees, brand allocation, posting and the way we plan and buy TV. It doesn't leave much of the old town standing.

A TV optimizer can quickly identify the combination of spots that will optimize reach or cost. The job is to optimize, not maximize, because the best solution can be either highest (as in reach) or lowest (as in cost). It's done with real Nielsen respondent data, not by formula.


The fundamental question is: "What to optimize?" The brave answer is "sales." So we might ask of the perfect optimizer, "Take this media budget and spend it so it will ultimately produce the most sales for my brand." The perfect optimizer -- who is not stupid -- might reply, "OK. But first tell me how advertising works to produce sales so I can devise a media strategy for achieving it."

That is where media planning needs to start but seldom does. How does advertising produce sales? Current thinking is it works by influencing the brand selection of consumers ready to buy the product. This is recency planning, and it says optimize reach.

So the second question for an optimizer is: "Do all reach points have the same value?" The answer is a resounding no. Reach points have a higher value if they add difficult-to-reach viewers. The optimization process finds hard-to-reach viewers far better than we do now. But reach points also have a higher value if the viewers are more likely to see and respond to commercials. Optimizers don't consider this kind of value unless ordered to. This is a big change.


Today we plan by daypart to reflect the different values of prime, day, cable and so on. Low-cost day does not compete directly with high-cost prime because planners assign them different values and functions -- prime for reach, day for frequency; prime for impact, day for continuity. Assigning budgets by daypart protects higher-cost TV.

Optimizers do not budget by daypart. They consider all dayparts interchangeable, which makes lowest CPM the principal driver. Optimizers often start and end with low-cost TV.

If we believe there are different grades of TV -- that a viewer who sees our message in "Seinfeld" will be more likely to respond than the same viewer who sees our same message in "Ricki Lake" -- then some kind of value weighting is necessary. If we optimize on CPM-reach alone, it is certain we will not optimize on effective reach. That is what I call the "Midas problem" -- turning everything into CPM.

All media price on qualities independent of audience size; i.e., the sight, sound and motion of TV. These other attributes are often more important than CPM in selecting media types. But media planners are uncertain how to assign values within TV. We act as if they exist when we select dayparts and programs, but since they've never been measured rigorously they are hard to defend.


Our current thinking focuses on attention as the mediator of a message effect. Viewer attention to what's on the screen has a direct bearing on whether commercials are seen. And we have a shot at measuring attention.

Probability of seeing the message can be modeled from information we have right now. DDB Needham Worldwide and Ogilvy & Mather are exploring viewer loyalty. Work done by John Mc-Sherry of True North Communications suggests that the percent of program viewed may be an indicator of attention to commercials. Our goal should be to use this data, as well as CPM, to evaluate TV buys.

In today's TV markets, what you buy has a greater bearing on cost effectiveness than how well you buy it. Since the reach value of a spot is in large part a function of its cost, the real market price is critical to optimizing. This means buying should drive planning -- the opposite of what happens now. The current practice of building next year's TV plan with this year's cost experience is bad practice. As the buy is being negotiated, changes in daypart pricing must be allowed to change the optimum scheduling mix.


Since optimization is based on reach, the planning unit for optimization should be weekly reach points, not ratings points. A $20 million brand TV plan might be written as "buy as many weeks as possible at a 38 weekly reach." There would be no daypart distribution, weekly target points or total target points indicated in the plan. This is very different from the straitjacket plans -- 25% prime, 25% day, 20% cable, 10% syndication -- planners give buyers to execute today.

Optimization is by brand, buying is by corporation. Individual brand optimization will have to be combined for corporate buying and then reallocated to each brand. Since reach is the optimization goal, cost-per-reach-point (not CPM) will be the allocation standard. The goal of allocation will be to achieve each brand's reach plan at the average brand cost-per-reach-point. This is different from the current target-point allocation model.

Optimization forces attention to another issue media people like to avoid -- the translation of a computer solution into an actual brand schedule appearing on TV. With reach optimization, enactment is hyper-critical because it requires specific spots to run in specific weeks. The tolerance for substitution or delay is far smaller than in traditional planning.


The enactment problem has been ignored in conventional TV planning and buying because of the guarantee system. In exchange for reduced risk, buyers have been willing to accept wholesale substitutions and less precise timing of audience delivery. Optimization will make these practices unacceptable.

Optimization programs are not new . . . [but] this second age of optimization is different. It has both the Pentium II processor and an urgency born of fragmentation. We need an optimizer to help buy TV today because we have so many options that just sorting through the inventory is beyond current systems. What we learn by optimizing will change the way agencies plan and buy media.

Mr. Ephron is a partner in Ephron, Papazian & Ephron, New York.

In this article:
Most Popular