For starters, you don't always get what you pay for -- even on data-rich sites like Facebook.
Want to find an audience online and then effectively engage and sell to it? One of the standard recommendations has become behavioral marketing—largely made by the companies that provide services and products in the field. The pitch is obvious: You want to do some advertising and a sales rep points out how much more effective paying extra for better targeting will be.
To be sure, knowing more about prospects before you send them marketing messages will almost always improve your results.
Unfortunately, many of the promises about just how well companies can target potential customers online are based on myths. Here are some of the bigger ones you'll hear:
This is the marketing equivalent of clothes making the person. By looking at someone's interests, a company is supposed to know who she is. But that's silly. All you know are her interests. That's important, sure, but to assume that you can infer more about the identity is dangerous. Look at this example of mis-identification from The Next Web. Google identified the site's Middle East editor, Nancy Messieh, as a male between 25 and 34 years of age. The age range was right, but male? Clearly off the mark.
I checked what Google had for me on Chrome on both my desktop and laptop: Gender was correct, but age off by at least a decade. Ah, the powers of immaturity. But that shows the problem. Assuming you can determine age via a collection of interests as demonstrated by websites visited means you have to buy into the research assumptions of the advertising medium.
And what if you run multiple browsers, some of which you run in stealth mode and some not? On Firefox, Google has no demographics or interests associated with me because there is no cookie. If multiple people use the same machine, the insights will be even more dubious.
Myth #2: Keywords reveal interests
Go onto Facebook, type a status, comment on someone else's post, or send a private message. Chances are you'll see an ad appear based on keywords you just typed. However, there is a slight problem. Although keywords can be telling in some cases, taking them out of context can produce the most ludicrous of results.
There's anecdotal evidence aplenty. Look at your own experience. Ask people on Facebook. You'll find many wonder why they get some types of ads that bear no relationship to their interests. The reason is clear. Words have multiple meanings, and unless you have a system that has the ability to detect the context (something that good search engines or smartphone voice recognition systems like Siri on the iPhone often do reasonably well, and even then things can go very wrong), you'll get many wrong implications. Just because you thought some company "made a hash of a situation" doesn't mean that you're looking for breakfast.
Beyond anecdotes, there is more quantitative data, like the study reported by Mashable last year that pegged Facebook ads as half as effective as banner ads. Talk about missing a low bar. Some argue that Facebook gets this result because of all the other compelling content, but you could probably say the same about many content-oriented sites with banner ads.
Myth #3: Find the interests and you'll find the buyer
Another classic mistake explained well in the blog of audience targeting company Semcasting is that interest is different from ability. Someone who dreams of travel and keeps looking at a vacation package to a Pacific isle may not have a cent to his or her name. Then again, perhaps the person has the money but puts it into a child's college fund. There's a difference between having an interest in a product or service and actually being able and willing to afford it.
Myth #4: You have to take it on faith
Many in the behavioral marketing industry want you to become a believer. They make general statements that sound reasonable. How can knowing what people do online not help your marketing? And it might. Then again, as the examples above show, it also might not.
Happily, this is the easiest myth to overcome. It's digital technology, folks, so consider trying behavioral marketing techniques, but in a test. Get the provider to run an A/B split test where half the appearances of an ad or delivery of an email is with behavioral targeting and half is with a run of outlet, where the messages could appear to anyone. Look at the response or conversion rates and see if they justify the extra cost of the behavioral targeting. If so, then you can continue the experiments, testing one type of demographic or interest factor against another and refining your program over time. If not, save the extra money and put it into marketing that makes a difference.