"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today's column is written by James G. Brooks Jr., founder and CEO atGlassView.

Digital marketing for the last few years has focused on wringing every last atom of efficiency out of ad buys.

This is a laudable goal in theory, but in practice it's misguided. In particular, it demonizes impressions that are aimed at consumers who won't ever buy the advertised product or service.

Of course, the other type of wasted impression – fraud – is a legitimate concern.

The industry has often gone so overboard trying to weed out wasteful buys that it has missed the big picture: Overtargeting isn't necessarily a waste. Exposing a message to a broad group of people is likely to have positive effects. If advertisers didn't believe this, no one would pay upward of $5 million for a 30-second Super Bowl ad.

But this approach makes sense for several reasons.

People outside your demo talk to people, for example, in your target demo. More than 80% of Americans seek referrals when making a purchase of any kind. Some 67% of consumers say they are more likely to purchase a product after a friend or family member shared it via social media or email.

Another possibility is that ads may also be seen by journalists. That's good news because many people learn about new brands from articles in the mainstream press. Others learn about new brands from search engine results and comments on message boards, both of which are influenced by the quantity of conversations about a brand. Very narrowly targeting a campaign risks the opportunity to create buzz.

Broad buys are often less expensive than more targeted buys. It's a Catch-22: The ostensible reason for granularly targeted buys is that they produce high ROI. Yet the "I” in that equation – investment – rises as the buy becomes more granular and the amount of inventory shrinks. What's better? Spending $1 million to target 10 million people, of whom only 500,000 will be interested, or $1 million for 300,000 targeted consumers? That is often the choice since the more granular the targeting and the more variables thrown in, the more expensive such targeting becomes.

If you follow the logic of granular targeting, it eventually leads to zero. As Joel Nierman once argued, if you keep cutting away the half of your advertising that's not working, eventually everyone in the target is already a customer or in the "not working” half.

Even tech giants like Google and Amazon still do broad-based TV campaigns. They have the means to run more targeted advertising with their own data but choose to disseminate their messages broadly instead.

And finally, brands can catch on with unintended audiences. In theory, targeting just the likely audience will create a group of early adopters who influence the rest of the market. But this vision doesn't take into account all the brands that have caught on with unintended audiences.

For instance, Timberland became an unexpected hit with rappers. Research has shown that people who engage with Santa Claus ads on the internet were 355 times more likely to respond to ads for nicotine gum than average. And people who liked "Mad Men” had a strong affinity for Williams-Sonoma products. These are correlations you wouldn't discover unless you executed a broad buy.

This is a great age for marketers. Advertisers have a lot more options than in the past. But they shouldn't be overly dazzled by granular targeting. When it comes to raising awareness, more can sometimes be better.

Follow GlassView (@GlassViewMedia) and AdExchanger (@adexchanger) on Twitter.