Melissa: Thanks for participating in our third eChat. Today we'll be discussing behavioral targeting. Let me start by asking any of you: How do you use behavioral targeting? Can you offer some examples?
John-Scott: Shopping, degree of commitment, discount orientation, affinity. These are the types of behavior we are able to detect and respond relevantly to. Relevance increases conversion.
Toffer: Behavioral targeting is at the core of what ChoiceStream does. It's how we make personalized product recommendations for our customers like Overstock and Blockbuster. The beauty of the Web is that it provides a wealth of information about how a user is behaving. If you can use that information in real time to deliver personalized promotions, you can increase average revenue per session, total sales and repeat visits.
Ron: Our business model and technology enables us to track shopping behavior. Since we also have the user's home address as part of the experience, we are able to add location and shopping behavior to create a valuable targeted marketing opportunity.
Sarah: It also enables advertisers to reach audiences of interest beyond limited, and often expensive, contextually relevant inventory, thereby increasing frequency and the potential for cost efficiency.
John-Scott: We do things a little differently. As an example, let's say you run recruiting for Arizona State University located in beautiful, sunny Tempe, Ariz. Imagine a prospective student is just about to arrive at your Web site and you detect that they are from St. Louis, Miss., that they have looked at the Web sites for the University of Michigan, Creighton [University] and Purdue [University]. For each of those institutions, they navigated to the Fine Arts curriculum section. You also noticed that they had viewed both the national Delta Chi and Sigma Chi Fraternity Web sites. So, we know this prospective student is probably male, interested in the Fine Arts program and just might be interested in warmer weather!
Ron: Another unique attribute that RecycleBank provides is that we have a long-term contract with the user's city, so we capture our data over many years and can change and develop our marketing message with the growth and changing lifestyle of the user/household.
Melissa: When you talk about behavioral targeting, are you talking about using it for e-mail? Online display ads? Search? All of the above?
Sarah: At Mindset Media, we actually look beyond the face value of the behavioral information and examine total clickstream patterns that indicate someone's mind-set, or psychographic makeup.
Toffer: We use it to help our customers put personalized product recommendations on their Web sites, but also leverage e-mail, TV and mobile phones as customer touchpoints where personalized recommendations can be delivered.
John-Scott: We wouldn't welcome them with the same old one-size-fits-all Web site - it could also be used to change online display ads if you wanted to.
Sarah: We use it to amass large audiences of people that share personality traits that fit a particular brand or category. Like 2 million people who are very pugnacious (and more likely to own or want to own a truck).
John-Scott: Ahh, great Sarah! So we would work with Mindset Media to make the visitor's experience even more relevant -- the right truck models to match the purchase ecologies.
Jere: At Prospectiv, we use behavioral targeting to upsell and cross-sell our client offers based on both implied and implicit data points we have on our traffic.
Ron: At RecycleBank, we [focus on] quality over quantity. We don't sell on the basis of number of page views. We sell on the basis of the deep and long-term relationship that we have with the user/household and the fact that they are transacting (in "spending money mode") when they are on our Web site and viewing their account.
John-Scott: Just to be clear -- we help our clients respond relevantly to their market segments (semantic personas) upon arrival to their Web sites. [Semantic personas form the basis of ThoughtLava's Semanticator marketing technology.]
Melissa: Can any of you discuss some trends that are taking place in behavioral targeting right now?
Sarah: One important issue is certainly privacy protection. Consumers are increasingly concerned that personally identifiable information is mingled with nonpersonally identifiable information, which they fear might be used to discriminate against them at some point in the future.
John-Scott: Well, there is the backlash in consumer privacy concerns.
Toffer: One general trend we see is our customers in retail and media and entertainment increasingly using behavioral targeting to deliver truly one-to-one, personalized recommendations or promotions versus traditional cohort-based, less granular targeting.
Jere: Yes, I agree with Toffer. This is what I am seeing, as well.
Melissa: Getting to the privacy stuff. How do you put consumers' fears to rest?
Toffer: By delivering personalized merchandising and marketing based on anonymized data and behavioral patterns.
John-Scott: We have found that when you gather behavioral data with personally identifiable information, you might draw the attention of the privacy watchdogs. Even worse, if you decide to share this information with others.
Jere: Agree with John. Very important that you don't share the data with others. That's a big no-no.
Sarah: Yes, we have seen that, as well, and believe that the industry is just scratching the surface on the interesting ways nonpersonally identifiable information can be used to define audiences of interest.
Ron: 1. We have the user sign a privacy statement. 2. We are transparent as a company. 3. We take privacy concerns seriously. 4. We never provide personal data to others.
Sarah: Transparency is the antidote. Privacy policies written in English, easy opt-out programs, clear distinctions between PII [personally identifiable information] and non-PII [nonpersonally identifiable information]. The former is clearly anything that can be used to identify you (e-mail, Social Security number). Right now, consumers are not very clear on the distinctions and fear that observation of their Web traffic (anonymously) can somehow be linked back to them as individuals.
John-Scott: It really comes down to returning value;privacy becomes an issue when there is no value, and they feel exposed.
Melissa: If someone was getting started with behavioral targeting today, what would be some best practices or tips you'd tell them?
John-Scott: Start yesterday! No, just kidding! They need to think through their market segments [and figure out] which are most likely to visit their Web site today. Once they have them in their minds, if they knew they were there, how would they respond to them?
Toffer: I think the important thing is for people to be clear about their goals to start with. Behavioral targeting is a tool that can be used to many ends.
Sarah: I was writing exactly the same thing, Toffer.
Toffer: Great minds ... ?
Jere: And expect to do a lot of data analysis, both before implementing behavioral targeting and after. You are never done modifying your behavioral system, and you must dive into the data to know if you are making the correct improvements.
Ron: Find a way to provide unique value or capture a unique set of data attributes. People know that this is where the industry needs to go. The key is owning, dominating or being the best in a segment of the industry.
Sarah: In terms of goals, you would approach [behavioral targeting] solutions very differently if you need to generate small lists of buy-now prospects versus seed awareness/affinity among prospects.
John-Scott: The goal is pretty universal. If you increase relevance or meaning (to use the semantic term), you will increase conversion. If you increase conversion, you will increase sales!
Melissa: Can any of you point to some great examples of behavioral targeting that you have seen?
John-Scott: We absolutely believe in the value of analysis. However, if we spend a little time thinking about visitors, we can quickly generate relevant content for the kinds of people visiting a Web site and respond relevantly. In one case, we were able to understand the urgency of a visitor by establishing their location, day of week and time of day. So when they arrived on our client's automotive site at 8 p.m. on a Friday, we could incent them to take a Saturday test drive in exchange for a gift card. Our client liked the idea because it limited exposure to only those prospects that had already decided to visit on a Friday and who matched a specific persona. We also could turn the offer off after a set number of prospects had accepted the offer!
Jere: Great [behavioral targeting] is done in such a way that people don't notice it, but the content makes sense. John's example follows this precisely.
Toffer: Blockbuster.com delivers personalized DVD recommendations on every page of their site based on rental history, movie ratings, etc. -- a pervasive approach to behavioral targeting.
Sarah: We think the next frontier is predictive targeting. Rather than picking off individuals who have "tripped a trigger," indicating they are in the market for something, and attempting to influence their decision at or close to the point of purchase, the next generation of targeting will focus on helping marketers "fill the funnel" with more of the right people before they register interest.
John-Scott: We're just trying to make our Web sites more intelligent and behave like the best salespeople/company representatives to return real value to Web site visitors.