Horizon Media’s e-commerce unit Night Market is launching an artificial intelligence platform that plans and optimizes investments across retail media networks and allows for in-campaign changes to shift money among networks, products and tactics based on predicted results.
Named Neon, the new platform is said to be able to increase advertisers’ revenue by 20% or more, according to Randy Browning, president of Night Market.
According to Night Market, Neon draws on historical media mix modeling (MMM) return on investment reports, client sales data and retailer-reported campaign performance data about return on ad spending (ROAS) to predict quarterly ROI. The platform predicts optimal media investment across retailers and third-party channels such as Amazon and Walmart, including breakdown between formats and tactics such as search, display and promotional offers. And it optimizes spending in real time by automatically predicting thousands of “micro optimizations” across retailers, tactics or within specific tactics to improve ROI.
Hershey Co. is an early user and endorser. “We have been thoroughly impressed by Night Market’s Neon suite of tools and the potential to accelerate our business in retail media measurement and product merchandising,” said Eric Bowers, senior director of omnichannel marketing at Hershey, in a statement.
Retail media and e-commerce have become growing focus areas for big agency holding companies in recent years, as seen by the launch of Omnicom Media Group’s Transact unit; Publicis Groupe’s acquisitions of Profitero and CitrusAD; and WPP’s acquisition of a series of e-commerce agencies and consultancies.
Independent Horizon’s move with the Neon platform—which is most readily applicable to packaged goods marketers—is more surprising given that it hasn’t been seen as a major mover in e-commerce and retail media up to now, and since its only major CPG client is Hershey, said one agency executive.
But Browning said Night Market also sees interest in Neon among non-CPG clients, including in retail, health and wellness and financial services, and the unit has been developing its capability since its launch almost three years ago.
Rather than make an acquisition or a partnership, Night Market decided to try more of a “white board solution” to build what marketers need from the ground up, Browning said, asking “what’s that end state we all strive to have, and let’s figure out how to get there.”
“The retail media space is like a wild west at the moment with all the money coming in,” Browning said. “And media is somewhat of a black box with each of these retailers. Therefore brands can’t bring the kind of best practices we know so well in digital media to the retail world.”
That means trying to apply the same media planning and evaluation approaches across retail media as marketers apply to other verticals, he said, even though the space works quite differently, often with people doing the buying assigned to manage brands within specific retail accounts rather than across multiple accounts.
Certainly, it makes things harder that different retailers have different rules on data sharing and report sales data and ROAS differently. For example, Walmart prohibits brands from sending sales data from its Luminate data platform to agencies or other third parties to mingle with sales data from other retailers, according to people familiar with the matter.
But in a statement, Night Market said Neon is set up to “operate behind our clients’ firewalls to enable sensitive data analysis.” Alternately, “If a client wants Night Market to run Neon, we will fill sensitive data gaps with customized retailer-specific machine learning models that learn the patterns between retailer shared data and third-party data to infer true sales, which we have found to be highly accurate within a significant confidence interval.”
Browning said Neon also is designed to help brand marketers try to achieve unduplicated reach and optimize spending against the same consumers across multiple retailers where they shop, using cohort analysis when, as is usually the case, it’s impossible to anonymously trace the same shopper across retailers.
At the broadest level, Neon is designed to help marketers plan a couple of ways, Browning said. They can start with a fixed retail media budget and use the tool to allocate spending to maximize revenue. Or they can start with a revenue goal and use Neon to estimate the retail media outlay necessary to meet it, he said.
“It gives brands a data-driven reality to how they plan,” Browning said. And it also can help CMOs move from their area being seen as a cost center toward being seen as a revenue driver.
“It’s not just a cute media tool,” he said. “It has an operational impact.”
Neon is also designed to analyze decisions at the individual product display page level across retailers, he said, conceivably leading brands to shift spending from the same item from Amazon to Walmart or vice versa, sometimes based on competitor performance at those retailers. For products and retailers that require a specific threshold for free shipping, the system may also suggest a product to co-promote to help customers hit the number, such as a dog toy that will trigger free shipping for a purchase of dog food.
