Matthew Liebmann, Global President of Movio Cinema, recently discussed the importance of behavioral data in loyalty marketing with Loyalty360. Check out his thoughts on personalization, demographic vs behavioral data and the future of loyalty below.
Can you talk about the difference between demographic data and behavioral targeting and why the latter is more impactful?
Demographics capture the characteristics of a group of people, such as their age, gender and ethnicity. Historically, this approach has been a useful proxy for targeting people in the absence of other information, i.e. people who look like this, tend to do that. However, it is assumption-based.
Behavioral data allows people to be targeted based on what they do, such as how they spend their time and money. This form of data looks below the surface to what a person’s preferences are and what motivates them to act. As the cliché goes, 'Never judge a book by its cover'. When you rely solely on the ‘packaging’, you miss the depth that lies underneath.
Can you talk about how a customer behaves, and that pattern of behavior, serves a brand very well as opposed to just knowing who they are demographically?
Let me give you an example from the movie industry, which is where Movio operates. Florence Foster Jenkins, starring Meryl Streep, is the true story of an American socialite in the 1940s who fancied herself as an outstanding opera singer when, in fact, she shouldn’t even sing in the shower.
Everything about this movie suggests that the demographic target is Caucasian women aged 50+, and this is largely borne out in the data. However, while 52% of the movie’s audience are women aged 50+, that still leaves almost one in two who sat outside of this core demographic. In fact, one in ten people who saw Florence were aged under 40. That’s a huge amount of potential guests and revenue left on the table by relying solely on demographic assumptions.
That’s why we’re fond of saying, “We don’t care if you’re an 80-year old woman seeing Teenage Mutant Ninja Turtles or a teenage boy seeing Philomena; it’s how you spend your time and money that counts!”
What are the key challenges to effective behavioral targeting?
The key challenge to behavioral targeting is collecting, and being able to use, comprehensive data. While data collection may be relatively easy in an online world - tracking what website people visit, what links they click and items they buy – it is much harder in the real world where many people still pay for their purchases with cash. In such cases, one cannot simply extrapolate online behavior to what is occurring offline.
Let me give another cinema-related example that illustrates the risk of extrapolation. Many theaters offer online ticketing now. However, online tickets represent less than one in four sold across all movies over the course of a year in the United States. A key motivation for pre-buying tickets online is fear of missing out, otherwise the majority of people prefer not to pay the online booking fee when seats will be readily available when they arrive at the theater. This means that online ticket data is heavily weighted to the opening days of movies like Star Wars: The Last Jedi, and not week five of The Dark Tower. In other words, it presents only part of the picture - which can lead to incorrect assumptions about someone’s likely behavior based on a subset of their activity.
Can you talk about the challenge of personalization today, a lofty goal for most loyalty marketers?
Personalization is more than tailoring a message to an individual by using their name in the copy. It extends to understanding the context in which to serve that ad for each individual recipient. A study conducted last year determined that purchase intent increased by 13% and recommendation intent by 11% when the ad was contextually served. In a world of programmatic advertising, it can be hard to guarantee that your message will not only be served in the right context, but not in one that actually detracts from your brand.
In addition, personalizing a message also means knowing when to stop sending it. How often have you searched for a pair of shoes online, ultimately purchased them somewhere but still had ads for the shows follow you all over the web? In order to suppress messages that are no longer relevant because the item has been purchased, advertisers need to be able to close the loop between message exposure and the transaction, wherever and however that transaction occurred.
Finally, whilst data provides enormous opportunities to personalize messages, there is a fine line between relevance and creepiness. If I search for shoes online and see ads for them on other websites, I know why I am seeing that ad. When subtler and more passive means of data collection are used to present ads without an obvious connection, heavy personalization can feel very ‘Big Brother-ish’ to some, and may lead to an unintended backlash.
What do you see as the future of behavioral targeting and how will that impact customer loyalty?
The future of behavioral targeting will involve predicting what a person is likely to do next in near-real time using machine learning – a form of artificial intelligence in which computers imitate human thought, in a way which leads to continuous improvement in activity and outcomes.
This approach allows marketers to identify trends and make predictions based on a consumer’s propensity on a depth and scale that is near-impossible for a normally-resourced marketing to team to undertake. It then allows messages to be served at a time and in a manner which continuously optimizes their effectiveness.
It is likely that behavioral targeting will evolve to encompass a person’s mood. So, not only who I am, what I like and where I am but how I feel at the time. As an indication of how things might progress, last year Apple bought a company which produces face-based emotional recognition products, and it previously applied for patents for mood-sensing technology, all with an aim to provide more compellingly-relevant advertising.