Commentary

How To Fix (Or Avoid) Bad Targeting

A Notre Dame fan recently visited ndnation.com to read up on the hard loss of the national title. Upon his visit, he was welcomed by banner ads displaying "Congrats Alabama State Champs." Upon clicking through the adchoices icon, it became clear that Google was behind the ad targeting. So why would Alabama ads be plastered all over a Notre Dame Web site? The answer is simply bad targeting.

One often hears that the solution to too much detail is automation and computerization -- simply allow the machines to take over. And there's no doubt that without the massive amount of computational power available to us today, we wouldn't have real-time bidding and an abundance of other sexy advertising tools. But as some clever person once said: "A computer is only as smart as the person programming it." And any individual working in modern ad operations will tell you, not everything can be automated.

Google is the most successful advertising company in history and one of the largest companies in the world. However, Google's current mantra is to automate everything. They simply don't have time to look over every ad campaign that's running throughout their systems. In this example, the page featuring the ad discussed football and Alabama -- which might lead to contextual targeting. The site visitor was also a football fan who had searched for information about Notre Dame specifically, which could have easily been a bad act of retargeting and behavioral advertising.

After reviewing the data points, it’s clear that this site visitor did not want to view an Alabama ad after his team went down in a 44-7 defeat. But the only real way to safeguard against this would have been to manually create a blacklist of sites for the ad.

As an industry, we tout better targeting, more precision and increased relevancy. Yet to do that, marketers need smart machines and a strong team behind them to determine how to accurately leverage the vast amount of data.

Here are three ways to avoid bad targeting:

  • 1.    Build campaigns around data; don’t build data around campaigns: Let the campaign strategy drive the data strategy, not the other way around. There is always a way to use more data.
  • 2.     Dig into your strategy: Identify targeting strategies and determine how they work together. This will reduce ad redundancy and help to avoid paying for ad impressions that you don’t need.
  • 3.     Make sure you can easily talk to your media partner: If you have trouble reaching them, or indeed if you have any communication issues, let this be a sign. If you can’t reach someone on the other end, then it is unlikely that a human is watching closely over your campaign.

Automation and big data play a role in campaigns, but again, we’re only as smart as the person programming the machine. When it comes to choosing a partner, marketers should look beyond the monoliths and consider those experts that do not completely rely on automation and big data.

Tags: commentary
Recommend (4) Print RSS