From Programmatic to Predictive Marketing: How Sizmek Harnesses AI to Optimize Brand-Customer Connections

TL; DR: Sizmek provides marketers and agencies with tools for negotiating the complexities of programmatic digital media buying. It accomplishes this by leveraging artificial intelligence to process the massive amounts of data necessary to assess individual consumer buying trends. Sizmek then acquires the most promising ad inventory and delivers customized messaging in real time to convert customers more efficiently. By using AI to optimize every stage of the customer journey, Sizmek increases the predictive power of programmatic marketing. And because its total transparency model measures the effectiveness of media campaigns against costs, clients understand precisely how Sizmek improves marketing ROI. From data enablement to creative optimization to media execution, Sizmek helps programmatic marketers use media inventory to its maximum potential.

More than 4 out of 5 digital display ads will transact programmatically by 2019, to the tune of almost $50 billion in ad spending, according to a report by eMarketer. As a demand-side ad buying platform, Sizmek bids for that programmatic ad inventory in real time on behalf of its clients. Transactions occur in milliseconds — faster than the blink of an eye — and the goal is to make every impression count.

In the rapidly evolving world of real-time programmatic digital ad buying, Sizmek leverages a powerful artificial intelligence platform to increase conversions and improve ROI for brands. In applying AI decisions across the customer journey, Sizmek reveals what data says about individual buying patterns, targets messages, and delivers them to the right content platforms at the right time to turn impressions into conversions.

“We use AI to uncover and act on previously hidden dimensions and connections between brands and their customers. Because of that, the experience delivered by each ad impression is enhanced,” said Nikos Acuña, Chief Visionary at Sizmek.

What makes Sizmek unique, Nikos said, is that it covers the entire journey of the ad, from conception to conversion.

Clients leverage Sizmek’s data enablement capacity to identify insights across five dimensions of predictive marketing — campaigns, consumers, context, creative, and cost. Armed with those insights, marketers can construct more intentional strategies for delivering inspired advertising.

Dynamic creative optimization tools and the world’s largest independent ad serving platform ensure those strategies can be enacted on a global scale. AI even helps Sizmek quickly change course in response to audience reaction and evolving campaign goals. The demand-side platform completes the picture.

“Media execution is about understanding who needs to be addressed, where to find those people, how much to pay to reach them, and what’s going to resonate with them,” Nikos said. “At Sizmek, AI powers it all. It’s a very robust and sophisticated platform that can act on precise insights all over the web, all over the world and all the time.”

Sizmek helps the brands it works with become more effective advertisers. As e-commerce continues to transform the retail industry and customer journeys grow more complex, Sizmek becomes a differentiating factor capable of building connections that can attract and retain business.

“We have refined models that help e-commerce marketers achieve their goals by optimizing to very specific outcomes,” Nikos said. “We look across all the available datasets and find the converters that work best. We show our clients who’s converting and when and what inventory they should be using to improve those numbers.”

Given the exponential growth of consumer and market data, however, those types of programmatic capabilities simply aren’t enough anymore. But Sizmek has the power to bring predictive value to programmatic marketing.

“Within the retail space, it’s about anticipating customer preferences now,” Nikos said. “We take programmatic to the next level by evaluating every single impression opportunity based on the individual’s propensity to convert and then putting money on the table — applying the optimal cost to each impression.”

The propensity score is Sizmek’s vehicle for accomplishing that empirically. Every impression up for bidding receives a score, and a lot goes into its creation.

“The most sophisticated brands understand the potential of the first-party data they have generated and have at their disposal. ”And then you bring in rich third-party data, the broader landscape of user preferences and behaviors garnered as individuals move across the web,” Nikos said. “Tying all of that together in a way that’s actionable — that’s what moves the needle. Because now you know not just what the customer has done in the past, but what that means for today and tomorrow.”

Having the information is a good start, but impression opportunities come and go in milliseconds, and there are a lot of eyeballs in the marketplace.

“Imagine the process of placing ads at scale with every available publisher, every impression opportunity available in real time. You’re looking at about 300 billion or so impressions across the web that have to be evaluated in terms of context, media type, audience, targeting restrictions, time, day of the week — not to mention the signals pulled from third-party metadata,” Nikos said. “Now you’re looking at about 44 septillion combinations of impressions to categories to channels to messages to audiences.”

Given those numbers, perhaps marketers should be satisfied that AI works at all. But Sizmek’s dynamic creative optimization engine ensures that all brand interactions are meaningful, wherever and whenever they occur. Moreover, Sizmek’s data-driven capabilities now help marketers tap into their customers’ psychological and creative dimensions and even into motivations more fundamental than that.

“Retail has been taking steps toward truly dynamic creative by using product-level codes to create relevant messages across locations and broad audiences,” Nikos said. “But we’re going beyond that to anticipate and deliver messages based on dimensions we’re just beginning to understand.”

These new capabilities can bring content to consumers based less on a few facts and more on a complete picture of them as a person.

“Can you target somebody based on their desire to change the world? Based on their wanting to take more control over their lives? Based on their desire to stand out, or to conform? We think we can — and we can deliver those messages to those audiences,” Nikos said.

With Sizmek bidding on an astronomical number of impressions in any given campaign and the viability of those bids up for grabs until almost the instant when the ads are served, clients might wonder if their money is being well spent. Sizmek’s total transparency model unlocks campaign intelligence across multiple marketing dimensions to make ROI numbers more meaningful.

“We find that as more data becomes available and the analytical tools at our disposal become more rigorous and sophisticated, unpacking the granularities, in terms of how client data is being used, has become a critical differentiator for us,” Nikos said.

That means looking deep within AI to communicate with clients which data attributes encompassed within a campaign are most effective. “Armed with that intelligence, marketers are empowered to work collaboratively with AI to make their ad spends more efficient and effective,” Nikos said.

Sizmek is working toward an actionable understanding of customer individuality on a massive scale. And it brings that information to programmatic advertising using AI to make real-time decisions on ad buying.

“We’re beginning to unlock the creative potential of all this data and understand when individuals are most receptive to messaging and what drives them to conversion,” Nikos said. “That’s possible because our AI is continually learning from the interactions it generates.”

The system assesses which data attributes — location, customer action, media type, and much more — create positive or negative effects on conversions and converges toward inventory scenarios that make campaign goals more achievable.

“In real-time programmatic environments, our AI, quite literally, puts money on the table and evaluates its own success,” Nikos said. “If we bid on a specific impression opportunity and the customer doesn’t convert, we move away from that scenario. We move toward the scenarios that produce the best performance.”

The results demonstrate why all of those split-second decisions are made and can even uncover new pockets of potential customers. And that means success building on success — which is the way forward for brands in the era of digital transformation.