“The Sell Sider” is a column written by the sell side of the digital media community.
Today’s column is written by Madhura Sengupta, director of ad product technology at Edmunds.com.
With the rapid evolution of the internet, consumers have started to face a critical problem: information overload.
As a result, many publishers have begun to personalize content to create a more compelling user experience. For example, there are “recommended for you” content widgets and navigation suggestions that are adapted on a per-person basis.
By the same token, publishers should create personalized advertising systems for their clients, both onsite and offsite through audience extension. Ultimately, personalized ad products will help marketers get closer to their ultimate goal of reaching the right person, at the right time, with the right message.
One technique a publisher can use to deploy hyperrelevant and personalized advertising is commissioning a technology vendor or investing in-house technology resources to run dynamic ads. For instance, a travel and entertainment publisher can use dynamic product advertising to showcase a carousel ad with different vacation and getaway packages based on a person’s inferred geolocation, usually via IP address. Geo information can be used to feature local promotions, storefronts and products reflecting local tastes, all of which will likely create a more relevant and personalized experience for the consumer.
To run dynamic ads, a publisher will likely need a technical team to support the creation of a product feed and dynamic templates, as well as the deployment of a pixel to collect and analyze audience data.
A product feed, or list of products, is a structured data file that contains specific attributes used to generate an ad, such as a product name, description, landing page URL, image URL, availability or price. For example, the travel publisher can create a product feed with all of its hotels (image, description, prices), tourist attractions and potential incentives. This product feed can then be uploaded into an advertising system or platform, typically via API, to auto-create hundreds of ads on the fly.
The raw data from the product feed can then be inputted into a dynamic template to produce multiple ad creative variants. While the components and messaging of an ad template always rely on piecing together fields from the product feed, they can take shape in different forms to boost user engagement.
For example, an incentive promotion for reduced hotel rates may have a larger image with a “redeem now” call-to-action button. Different hotel amenities can be showcased in a carousel format to tell a story to users. Hotel room images may even be stitched together as a slideshow to create an eye-catching and exclusive offer.
Discrete templates can also be created for various hotel types. For instance, a luxury template showcasing a five-star hotel may emphasize premium and exclusive features, while a template for low-budget hostels may focus on free walking tours to appeal to student travelers.
After ad templates are created with a product feed, publishers can then implement a pixel to track user activity and determine which ad to show which user. For example, if a user has demonstrated an interest in scuba diving on a travel website, this information can be passed via a pixel event and triggered to promote the appropriate products in a product feed, such as tours and rental equipment in the correct geolocation.
Taking it a bit further, if a user has already booked a hotel from this website, the pixel can exclude them from any ad campaigns since they have presumably already “converted” on a purchase action. As such, intent signals from a pixel can be used to deliver the most relevant ads to the right person.
To get the most out of these ads, publishers can use audience targeting beyond just the pixel’s signals to improve performance and relevance. Data management platforms can provide additional layers, such as geotargets (radius-based, ZIP code or region), demographics and behavioral prospecting (interest in specific product categories), among others.
Machine learning can also be used to understand which creative works best with a given audience and automatically substitute the most successful variants for that audience. Advanced algorithms can analyze past creative performance to learn over time and adapt and conduct further substitutions should preferences change.
While configuring dynamic ads may be a large technical investment whether it is done in-house or with a vendor, this approach can create huge benefits for both marketers and consumers. Dynamic ad automation not only increases time savings for creative assembly, but also achieves more effectiveness in reaching the right person at the right time.
Since personalization and relevance apply across the board, almost every industry – from automotive to retail, entertainment, real estate and more – can take advantage of this technology to drive sales and boost revenue.
Follow Edmunds (@edmunds) and AdExchanger (@adexchanger) on Twitter.