• by Michael Kubin , November 7, 2000

By Michael Kubin, Co-CEO, Leading Web Advertisers

In the 1500s the use of early telescopes prompted fascinating discoveries. It became clear to the astronomers of the day that the universe demonstrated a remarkable predictability; heavenly bodies were shown to behave with the accuracy of a finely tuned timepiece. And even with the rudimentary tools of the day, this inexorable predictability soon generated a philosophical argument that survives to this day.

Tycho Brahe, a Danish astronomer, spent his entire lifetime assembling data on the heavens. His observations on the movements of planets and stars later permitted the great theorists to test and occasionally prove their hypotheses. This led to the mechanical model of the universe, an argument that stated that since all bits of matter are governed by the same laws of physics, all one needed was enough data and enough computing power and one could predict the future.

This was a safe argument to make at a time when neither data nor computing power were anywhere near sufficient to the task. Which is what makes this such a compelling theory: its consequences.

Think about it: if you could quantify the position of every piece of matter in the universe, and if you could compute their relative physical effects on one another (gravity, magnetism, nuclear forces and such), you could calculate where those bits of matter would be on, say, Saturday, and you’d know the result of the sixth race at Belmont. Or the gender of your next child.

Or the last word in this column.

A similar argument exists within media planning and buying.

After all, the media universe is mostly a business of numbers, with audience, circulation, and pricing following predictable orbits. The hyper-quantification and mechanization of how media is planned and bought is a perfect application of computers and their intelligent software extensions, optimizers.

Media mechanization is tempting on many levels: advertising agencies, operating on thin margins, continually look for ways to replaces humans with machines. And the performance pressure that advertisers live with leads them to look for optimal ways in which to allocate their media dollars. The huge amounts of data combined with the recent fantastic strides in computing power make for a tremendously appealing mix. And the Web, with its inherent ability to provide data on just about everything, is at greatest risk for providing advertisers with sterile, lifeless, computer-generated media buys.

Optimization engines combined with an ever-increasing number of comprehensive planning and buying resources, while efficient, could lead to a kind of media homogenization, with budgets being the principal determinant of visibility. But as opposed to the schoolchildren of Lake Wobegon, not all advertisers can be above average. In a world where all media plans were executed by machines, the ranking of media success would simply parallel dollars spent. How boring.

The big medi