The Non-Creepy Future of Personalized Marketing
By Mark Torrance
Ad-tech companies are finding ways to help marketers reach consumers in increasingly personal ways. By leveraging a history of activity and interaction with their brands, marketers and their ad-tech partners have the ability to really show relevant, personalized marketing messages that are tailor-made for each consumer.
Unfortunately, too often advertisers take a very clumsy, broad-brush approach that may leave consumers feeling misunderstood, unresponsive, and uninterested, or worse, creeped out.
For example, everyone has their favorite story about a product they browsed once on a retail website, where that product subsequently “followed them around the web” wherever they went. This reflects a particularly clumsy approach to retargeting. Yes, the consumer showed interest in that product, but why not focus instead on the brand or the store to try to get them back?
Why not feature other products that are best-sellers in the same category, or around the same price point, or complementary to the products the customer showed interest in? With these small changes, the advertising experience can shift from “creepy” to “understood.”
Mass-market retargeting is clumsy
In point of fact, this kind of rules-based approach can be a good start, and it’s much better than nothing. But when you work with a company that can apply predictive models to this problem, the results become even better. Models can factor in data and experiments across all past users and ad exposures to determine the optimal set of products, offer, and design elements (like whether or not to feature the price prominently). This approach solves the problem in a way much better than humans playing with a rules engine could.
Moreover, frequently these kinds of ads are really ugly and clumsy in their design. They look more like they were built by a programmer or a web-design student than by a professional creative-design team. This can happen when products from a diverse product catalog are expected to be fit into a “one size fits all” template, especially when the ad is designed to show more than one product in a product carousel.
Information can be beautiful
Instead, why not try something that is closer to the beautiful designs created by the top creative departments? With thought, the “carousel” concept can be reimagined to show large images of a series of products, gracefully cross-fading or animating within the ad to subtly call attention to them, and at the same time feature more of the content (the great design of the product itself) that will appeal to the customer. A beautiful presentation of the series of products offered can have a look that borders on native advertising.
Beware the arbitrary frequency cap
Additionally, overexposing customers to ads is not a good strategy, and can actually backfire. We find that for each campaign there’s an optimal frequency curve that shows us how many ads it is valuable to show to each customer, and after you reach the peak of the curve, showing more ads actually decreases the customer’s likelihood to buy. Even more challenging: the curve is not just different for each campaign, it can be different for each customer.
As such, frequency capping may do more harm than good for a given advertiser, as the slope of each customer’s curve may join to an average, but they are in fact very different. Would you not want to market to half of an interested customer base because of an arbitrary rule that doesn’t give your advertising the opportunity to impact that customer’s experience? Fortunately, by gathering lots of data, smart marketers can determine the shape of these curves, and tune their appetite for showing additional ads per customer so as to achieve optimal performance.
Winning them with context
Finally, dynamic creative ads are rarely served in a context that is highly relevant to the content of the ads. An ad for a retailer that features diverse categories of products can be many times more effective when it features products from categories that match with the contextual category of the page on which they are served. This kind of rapid, real-time contextual targeting is really tricky to get right, but it can be extremely powerful. At Rocket Fuel, we’ve seen performance improvements of up to 400% when adopting this approach vs. just featuring products based on what the consumer has previously shown interest in.
It’s time to move on from the first generation of tired, formulaic dynamic creative advertising, and to embrace a new world of elegant, subtle, thoughtful marketing that will speak volumes, and enhance the relationships marketers can build with their customers and prospects.
Mark Torrance is CTO of Rocket Fuel Inc.