How may companies tap into the explosive power of online behaviour targeting, while avoiding the pitfalls?
This and other issues were comprehensively addressed by Mladen Raickovic, manager, Olive Brand Response, OliveMedia, at the recent MESH Marketing Conference in Toronto.
OliveMedia is an online advertising consulting firm and ad network based in Toronto.
Behaviour targeting can dramatically enhance the effectiveness of an online ad campaign if properly done, said Raickovic, an expert in performance-based advertising.
Pull off a cookie coup
Behaviour targeting uses information collected on a person’s Web browsing practices – such as pages visited or searches made – to determine which ads to display or promos to offer.
Data is collected via tracking cookies dropped into the browsers of Web surfers.
READ RELATED ARTICLES
The expectation is properly targeted ads will evoke greater consumer interest, deepen brand awareness, and lead the targeted group to take desired actions, such as buying products, subscribing to certain services, and so on.
Some years ago, the technique helped American Airlines Inc. realize a whopping 115 per cent increase in business travelers making a minimum of one trip a year, and an equally astounding 145 per cent increase in travelers making more than five trips.
But achieving that level of success depends a lot upon how effectively – and ethically – behaviour targeting techniques are applied.
Cluster and conquer
Clustering your audience based on certain metrics is at the very heart of behaviour targeting, Raickovic said. It’s a capability offered by many of the systems available on the market today.
“Examples of typical clusters are: auto enthusiasts, female shoppers, in-market travelers, health seekers.”
He said these groupings are done based on information collected as users navigate through the Internet. “Their online behaviour and activity are analyzed and they’re slotted in certain segments.”
For instance, he said, a visitor to Zappos.com may reveal an interest in products such as shoes or blazers. “We may also learn about specific brands they are keen on – such as Gucci, Tommy Hilfiger or Ralph Lauren.”
Purchase intent, he said, is another piece of data that behaviour targeting tools discover by looking at what’s happening in the user’s shopping cart. “For instance, are they putting products into the cart but not following up right away? That information may reveal the person has purchase intent, but has held back for some reason.”
Frequent visits to sites such as Fashion Magazine or the fashion section of MSN would uncover content interests, in this case an interest in fashion.
Likewise, he said, visits to social networking sites such as MySpace or Facebook offer age or gender information. “It may tell us, for instance, that the person is a female. So assembling all these pieces of information, the system may put this person in the ‘Female Shoppers’ cluster.”
Raickovic also cited the example of someone who frequently visits sites such as Lonely Planet, Trip Advisor and WAYN (Where Are You Now). “As these sites are all specifically linked to travel content, it may indicate the individual is a travel enthusiast. Visits to Expedia.com and Travelocity to check out fares tell us he or she is in the market, either now or in the near future.”
Based on all this information, he said, the person may be classified as an “In Market Traveler.”
While focused clusters have their advantage, slicing and dicing your audience shouldn’t be carried to an extreme, Raickovic cautioned.
“For instance, in Canada we have 24.5 million people online. If an advertiser decided to target only men, the audience would go down to 12.3 million. If men between the ages of 25-34 years were selected, it would be 2.1 million. If it’s men, between 25-34 years, with an interest in fishing – the selection would go down to 250,000 people.”
He said clustering that got any more specific would risk losing its effectiveness.
“For instance, if you added another layer to this: men, between 25 –34, with an interest in fishing, living in Toronto, who’ve moved here in the past two weeks, you’re looking for a needle in a haystack.”
Objectives, Audience, Ethics, Partners are the four pillars that support effective behaviour targeting, Raickovic said.
“Have clear business goals and good reasons why you are targeting a particular audience,” the marketing expert said.
For instance, he said if general visitors to WalMart.ca are men and women between 20 and 70 years, living anywhere in Canada, it would hardly make sense for the company to do behavioural targeting because the audience is so broad.
“But if Wal-Mart Canada wanted to reach mothers in Toronto with a new line of products that would be a good reason to do a targeted campaign.”
Success metrics should be defined in advance, he said, so the campaign can be evaluated later. “Understand clearly what qualifies as success – it could someone buying a product, or signing up for a newsletter or joining your online community.”
It’s crucial that creative developed for the campaign actually speaks to the target group, Raickovic said. “You can’t target mothers and then have very broad creative. If you’re spending more money to focus on a very specific group make sure the message speaks to that group and there’s a strong call to action.”
An “actionable campaign”, he said, also extends to the design of the Web site. “So if you’re looking to get people to sign up for e-newsletters, drive visitors to the page where the sign up form exists, not your home page, or people will get frustrated and leave.”
While focusing on your target audience is great, Raickovic strongly urges marketers to build some flexibility into the campaign.
“Say your target audience is men and women between 25 and 34 years, interested in technology, with a high household income. It’s possible that be 75 per cent of your customers exist there. But by being a bit flexible you may be able to reach a whole separate pool of people.”
Campaign mangers, he said, should not be afraid of experimenting. “Running a successful campaign can take several iterations before you get it right so experiment with audience and creative.”
Avoid cookie-ing up a storm
While most marketers focus on results, when doing behavioural targeting you neglect “ethics” at your own peril, Raickovic cautioned.
“The last thing you want is to be on the wrong end of a privacy lawsuit, especially today when it’s such a hot topic.”
In a behavioural targeting campaign, your partners can either be your launch pad to success or drag you down to failure, Raickovic said.
He cautioned firms who use third parties for data collection to understand their information collection practices: where they get the data, how, what are they paying for it (“you want to know they’re not collecting it surreptitiously”).
Checking out the partner’s targeting capabilities is also crucial, he said.
“Some companies will offer you very basic segments based on limited amounts data collected. So if a person visits an automotive site just once, they’re classified as an automotive enthusiast. That’s not effective. You want to have multiple touch points for creating segment.”