Profiling for Advertising Effectiveness

Our series on the use of profiling in advertising has, like the email series, generated a lot of reaction from our readers. Our question a few weeks back was, “Does personalization in advertising work”? We noted that we’ve received a lot of feedback, both pro and con on this question. Now we’d like to share the comments.

First of all, it was fascinating to see how many readers are involved in businesses that relate to targeting and/or advertising efficiency. Many readers wrote to describe how their companies are tackling the challenge. A common theme in many of these approaches is that they rely on user-supplied opt-in information to drive their profiling systems versus assumptive profiling techniques based on user behavior.

One proponent of opt-in from WinWin Technologies noted that assumptive profiling:

    “…makes certain assumptions based on the contents of sites people visit. To some degree, these assumptions may be accurate, but we have no way of analyzing this, except to look at response rates but response is the result of many factors of which the correctly targeted audience is just one (product, offer, creative are some others). Would it not be much better if users voluntarily told us what their profiles were and what kinds of offers they were interested in?”

An opt-in system for banners? For example, if your client was Sony Vaio and you could target users who voluntarily told you that they were interested in buying a new laptop in the next six months, that they were PC users, that portability was important to them and that they lived in North America, would you not pay a premium to reach these users? The main difference between this situation and the former example is that the guesswork has been eliminated. These consumers have told you what their interests are, they have entered into a conversation, and have opted in to receive relevant advertising messages.

We sure can’t argue with how nice it would be to be able serve banners to people who have taken the time to opt in for them if, in fact, people will take the time to do that.

Our next reader also doubted that getting targeted banner ads would be sufficient incentive for people to provide personal information. This opt-in fan from ScoreCardUSA shared his thoughts on how to encourage opt-in so that effective voluntary profiling can be done. He notes that:

    “What typical personalization stories miss (and Amazon’s Jeff Bezos and others forget) is that in order to get true ‘personalization,’ companies need to motivate people to ‘volunteer’ meaningful personal information. For that to happen, customers need to see a significant tangible result in return. Do you see a real value in giving personal information to Amazon so that you can get ‘better-targeted ads for books’? Neither do I, and I am betting neither does any one else (except the advertisers of books). Virtually all companies are experts in some area (Weight Watchers, Home Depot, etc). People look to those companies not just for products and services, but advice. These companies need to proactively use their expertise (in dieting or home remodeling) and offer services and advice to their customers in a meaningful way. It is during that sincere exchange of information when customer loyalty and intimacy is built and can best be leveraged for the benefit of both the company and the consumer.”

Finally, rounding out the opt-in comments is another reader from who shared his company’s experience with registration forms:

    “Our experience shows that consumers are more than happy to complete a detailed registration form (ours takes about 2 to 5 minutes) if they feel they will receive a direct benefit for doing so. In our case, we use their registration information as the base for their demographic profile. We then use their demographic profile, shopping profile, and browsing profile to provide personalized, targeted, and customizable “admercials”… Target marketing based on a mixture of demographic, shopping, and browsing profiles is how advertising on the Internet was intended.”

Any fans of assumptive profiling for ad purposes out there? Tune in next week for more reader feedback. Want to discuss the merits of active versus passive profiling? Send us your thoughts, and we’ll continue this thread next week.

Related reading