Two California companies have invented technologies that quite literally create ads customized to the moments in which they're being viewed. This is a really cool and scary idea.On of the companies, Adisn, looks at page viewers, using a variety of behavioral and analytic tools to understand who's looking, what they're interested in, at which times, and where. The other, called Tumri, has developed technology that creates ads, pretty much in real-time, to address those likely tastes.

It's something out of a Philip K. Dick novel, isn't it (or a movie adaptation, like those billboards in "Minority Report" that morph to sell stuff specifically to Tom Cruise as he walks by)? The idea of outsourcing relevance and utility to a technology certainly feels futuristic, if not downright sexy, in a geek sort of way. Like the same way you could get excited about Pierre LaPlace's observation that we could predict the future if we knew the position of every atom in the Universe.

Knowledge should trump chance. If only a machine could accurately anticipate what people will do, the marketing business would be so much easier.

Only it can't. And the idea that is could is a little scary, isn't it?

Forget for a minute the unpredictability inherent in any observation at the underlying reality/atomic level (you can measure position or speed, but not both). People are even harder to read. Ads can certainly get configured with a higher likelihood of succeeding, but I'd wager that performance is going to get decided by, well, chance. Or call it fate.

That's because events are not identical or contiguous. Variables like individual personality, and unpredicted contextual elements, mean that relevance and utility aren't just matters of configuration, but ones of content and uniqueness. Saying something important, however imperfectly, might still be the more useful ad strategy than making sure saying something less important is otherwise presently perfectly.

There's also the issue of purpose. The best ads aren't the memorable ones, but rather the ones that get people to do things...like buy stuff, sooner versus or later. This isn't a technology challenge, it's an old-fashioned question of content: you just have to prompt a response. Whether clicking on a box, dialing a phone, or driving to a store, the behavior is technology-neutral.

There just needs to be a point to your ad. No tech required.

Adisn and Tumri have developed cool technologies, though, which might make ads work a lot better. They just won't ensure that they succeed, because people...and the future...remain at least somewhat unpredictable.

And maybe that's a good thing?

Jonathan Salem Baskin writes the Dim Bulb blog, and is the author of Branding Only Works On Cattle.