The Trade Desk Releases New AI Tools Meant to Open Up Black Box Programmatic

The company claims the platforms will cut down on unnecessary trial spending

The Trade Desk's new tools aim to make programmatic advertising more transparent. The Trade Desk
Headshot of Patrick Kulp

The Trade Desk is using artificial intelligence to give marketers a clearer picture of where their programmatic dollars actually go.

The ad platform announced three new tools on Tuesday meant to simplify the programmatic buying process and better predict who exactly advertisers can expect to reach ahead of each campaign. The rollout comes as brands have grown frustrated with the often-opaque nature of placing and measuring large-scale automated advertising results.

“We feel that we’ve revamped the user experience—we’ve adopted a new and transparent form of AI, and we think that this is going to make many of the complexities of programmatic much, much simpler for everyone,” said Kathleen Comer, The Trade Desk’s vp of client services. “We think that’s going to impact the future of programmatic because buying will be much more simple, more intuitive. And when buying becomes more simple and intuitive, it becomes more scalable, which is infinitely important.”

The backbone of the new offerings is Koa, an AI forecast engine that digs into the Trade Desk’s massive data trove to spot and analyze trends and patterns. Koa will power two other tools: Megagon, a user interface that presents the AI’s predictions and recommendations and offers a streamlined way to incorporate them into spending decisions, and The Trade Desk Planner, meant for mapping strategies across different devices and channels.

Comer said the planning tools eliminate the usual need for marketers to trial balloon ads with what’s known as a “learning budget.” Instead, the AI will continually learn from the results of previous campaigns across the platform to hone its ability to provide relevant metrics upfront before any money is spent.

“A lot of times, planning is more of an art than a science,” Comer said. “Plans have traditionally had to rely on proxies rather than data. We’re applying more data to something that was often guessed at.”

The company also made ease of use and clean design a top priority when building the user-facing layouts, Comer said. That emphasis was informed by complaints the company has heard from traders about the amount of time they are having to spend on troubleshooting their campaigns in unnecessarily complicated control panels.

Niklas Cederlöf, programmatic marketing manager at Berlin-based delivery company Foodora, said tools that incorporate as much outside data as possible are key when working with few internal resources in unfamiliar markets.

With a small in-house programmatic team responsible for more than 20 markets globally, Foodora “had limited resources when it came to optimizing these campaigns in order to reach our targets,” Cederlöf said in an email. “We needed an intelligent tool to easily help us launch campaigns in new markets.”

Open-exchange programmatic advertising’s reputation for nebulous measurement and limited control over placement has driven some brands to smaller private networks and other more regulated forms of ad-buying, especially in the wake of big media scandals over programmatic ads appearing next to offensive content in recent years. But spending on the practice has continued to grow at a steady clip, with programmatic budgets expected to swell nearly 24 percent this year, according to Emarketer. The Trade Desk is hoping new tools like these will enhance that growth by changing the medium’s perception in the industry.

“Sometimes programmatic can get knocked for feeling like a black box, but we do not believe in black boxes,” Comer said. “What we believe in is surfacing recommendations so that buyers understand how to better allocate spend and assign value.”

@patrickkulp Patrick Kulp is an emerging tech reporter at Adweek.