Commentary

Finding The "Answer" In Display Ad Targeting

Marketers who are dissatisfied with the generic demographics and data clusters that are widely used for targeting online ads to consumers should heed the advice of advertising veteran Bob Seelert, who advises marketers to begin with "the Answer." His approach is as relevant in today’s world of digital advertising as it has been in traditional advertising.

Seelert counsels marketers to decide exactly where they want to go, and then work backwards to determine the most direct route for getting there. When it comes to targeting ads to consumers on the Internet, the “Answer” is marketers’ own data. Think about it. Every marketer knows exactly who their best customers are. They know who represents the most lifetime value, whether it’s products or services being marketed.

So what’s the shortest path to finding new customers who look very much the same as existing ones? Behavioral targeting of online ads helps you find hand-raisers, but it’s not well- suited for honing in on your highest lifetime value prospects. Relying on demographic targeting or generic clusters is better than nothing, but these approaches are weak proxies for your best customer with many wasted impressions along the way. There is a wealth of offline consumer data against which to profile existing high-value customers and create “lookalike” targets for display ads.

Take a life insurance company as an example. It knows exactly who its most profitable customers are. Not only does it want to find more of them -- it doesn't want to waste time and money targeting consumers who don’t resemble its best customers.

The answer is not to purchase all of the offline consumer data and set out to mine it in search of the holy grail of targeting. After all, who buys the cow to get its milk? Instead, the company contracts with a data specialist who has access to many different data sources with which it can model more than 100 predictive factors that define a high-value life insurance prospect.

By combining the power of many data sets into a single audience model, this process can produce 10 to 20 million prospects who are most likely to become customers and generate above-average profitability.

This custom-built audience, derived from the marketer’s own first-party data, is then deployed using relatively inexpensive real-time bidding media inventory. Only consumers who match the high-value profile are served display ads, greatly reducing wasted impressions. Moreover, the life insurer doesn’t end up serving an ad for a $1 million policy to a prospect who can afford just $100,000 worth of coverage. The result of this precise audience segmentation is a scalable and cost-effective way to leverage display advertising.

The beauty of this approach is that consumers are targeted at the very top of the marketing funnel, before they’ve shown any “in-market” behavior for a particular product or service. By the time some of these prospects make it down the funnel to search channels, their search behavior has been influenced by the ads they were shown at the top of the funnel.

Every marketer knows there are lots of potential customers out there. The challenge is to find them cost-effectively. Start by examining existing customers. You already know what they look like. That's the most sensible answer.

Tags: commentary
Recommend (7) Print RSS