In the early days of online advertising, when many advertisers were using online media as a direct-response channel, testing at low volume across a wide variety of sites was a popular strategy.
Reps and agencies alike were big fans of the “throw a lot of stuff against the wall and see what sticks” approach. The overarching principle of this approach was that wide variances in response could be seen from site to site, and that even some sites that didn’t appear to be contextually relevant to a product or service could produce some surprising results.
But how relevant is this approach now that online advertising is no longer in its infancy?
We’ve learned quite a bit over the years about what makes Web users purchase online. We’ve learned so much that the spaghetti-against-the-wall approach seems silly when compared to other approaches that have their roots in sound marketing strategy.
Engage and other companies that profile online users have developed some interesting targeting options for those of us looking to reach Web users who might be interested in our offers. By observing surfing habits and applying recency, frequency, and duration algorithms, Engage can identify users who are more apt to respond to offers in a given category. By picking relevant categories of users, Web advertisers can abandon the spaghetti approach.
Certain agencies have been successful at leveraging past behavior data and aggregate client data to steer future initiatives. Past performance is usually a reliable predictor of future success, and certain assumptions about response can be carried over from the aggregate experiences of an agency’s client base. If you haven’t been using this data yet, the time to start is now.
We’ve also learned to target Web users by identifying interests that go hand in hand. If research suggests a strong positive correlation between buying a car and having bought a new home within the prior six months, it makes sense to test sites that cater to new homeowners if your client sells cars.
Past performance, profiling technology, and lifestyle modes represent only three of a number of assets that can be exploited in order to guide media selection for future buys. And they’re pretty powerful at boosting response. Why, then, do sales reps insist on submitting proposals for site- or network-wide tests? For one thing, it generates some revenue for the ad sales organization before rolling out a larger campaign.
It’s simply not enough to scatter your ad buy to the four winds and hope that something generates a higher response rate than you might have initially thought. That’s putting the cart before the horse. Truly effective online media plans are the result of rigorous research and innovative implementation of the various ways to reach your target online.
We’re now well beyond the “throw stuff against the wall and see what sticks” approach, so let’s focus on finding the target audience before advertising to them, rather than letting the tail wag the dog.