All the hype might make you think optimizers can magically increase the effectiveness of advertising and lower its cost. However, as more and more people experiment with the software, "optimizer-mania" has given way to a more realistic assessment. Optimizers can be useful tools, but not more. There is no magic -- it's only computer software.
NBC's evaluation of optimizers has led it to the same conclusion. We learned there are three essential elements to the successful use of optimizers: Know the optimizer; consider the limitations of the data; and apply the right advertising strategy. Under those conditions, optimizers can help. Otherwise, they are likely to do more harm than good.
KNOW YOUR OPTIMIZER
The exaggerated expectations that surrounded the introduction of optimizers into the U.S. market cannot be blamed on a new British invasion. Neither the Super Midas nor the X*pert representatives claim their software has revolutionized planning, buying and selling in Europe. In fact, we found them to be quite modest about what their software could do.
They even mentioned limitations, pointing out the TV audience data available here are neither as complete nor as inexpensive as European ratings data. They did say their systems, if used correctly, would be helpful dealing with the increasing complexity of planning and buying. One can't argue with that.
American optimizer-mania was probably caused by the magic words in their presentations -- efficiency, effective reach and, of course, optimization -- which made it easy to overlook the absence of the most important word: effectiveness.
Optimizer software cannot optimize effectiveness. It can help you implement strategies, but it cannot design a strategy for you.
Recognizing that optimizers are dependent on good planning is essential. However, there are other important factors to consider. Most optimizers were developed for European planners who work under different conditions than planners in this country. For example, there is little upfront buying in most European countries and most commercial pods are much longer than in network prime time in the U.S.
Therefore, an optimizer may be optimal for most European planners but, unless you modify it, it may do things that work against your strategy. Moreover, different optimizers rarely produce the same plans, even when given the same specifications.
POOR DATA, POOR RESULTS
The successful use of optimizers depends upon the availability of detailed and complete respondent-level data. While everyone complains about the price of the data, we are surprised by the apparent lack of concern about data issues that could significantly distort the results of optimization.
First, the unified samples used for the analyses in this country suffer from a much larger number of missing cases and missing data than in Europe. Since we know that respondents with missing data tend to be heavier viewers, the optimizer probably understates and distorts the true reach of the plan. Then if those data are compared to magazine reach models, which use rather generous reach curves, the distortions get magnified.
The lack of minute-by-minute ratings can cause further distortions. Having only one data point per quarter-hour can identify an "ER" viewer quite well. But if you are considering channels where most of the audience might be grazing somewhere else within a few minutes of the data point, the available data are just too imprecise for good planning.
Finally, if you optimize on program-level data and look for reach and efficiencies, you would be magnifying statistical error. When optimizers are told to look for cheap spots, they may select programs based on a sample of two viewers. The optimizer would not alert you to the large statistical error in that media plan. It deserves a red flag.
In short, you have to consider the implications of these data issues, understand how the optimizer deals with them and use good judgment.
STRATEGY IS EVERYTHING
Optimizers can optimize reach, frequency and efficiency, but there is no button for "effectiveness." You need a media planner with a Midas touch to set the qualitative criteria that will lead to the goal of efficiency.
It is true optimizers help planners in a complex environment. It is also true they are particularly attractive to planners who want to believe all GRPs are the same. But planners who must implement sophisticated strategies that focus on effectiveness face a difficult task.
We already know of agencies that have added effectiveness factors, such as viewing environment, attention and viewer loyalty, to the optimizer database. But these are qualitative factors and optimizers are quantitative tools. Incorporating attention, loyalty or the value of being in a program that 20 million people consider their "favorite TV show" into a media plan requires experience, expertise and judgment.
Consumer spending indices and attentiveness measures, as well as reach, frequency and recency, all point to broadcast networks as the most effective way to drive sales. Recently, an independent study presented at the Advertising Research Foundation's annual conference clearly stated broadcast network prime time is demonstrably the most effective daypart to drive sales. It is our job to show our clients we have the best way to reach the people who actually buy most of the goods and services in this country.
As in every other commercial context, optimization must focus on the value received, not just the cheapest media price. Value in this context is measured by audience behavior while watch- ing TV and when making purchase decisions. Because its audience pays greater attention for longer periods of time and spends more money on every product category, network broadcast TV is the best way to optimize value.
This, of course, is the advertiser's ultimate goal -- and therefore the only relevant goal of optimization.
Mr. Braun is president, NBC Television Network, New York.