Artificial Intelligence (AI), in all its forms, is among the most important and widely discussed topics in programmatic advertising, and for good reason.
While many media buying platforms are paddling frantically to catch the AI wave, we’re surfing ahead of the pack. We added AI capabilities to our platform back in 2016, with the introduction of KoaTM.
From the beginning, we’ve maintained that AI works best when it serves as a partner or co-pilot to you, the media buyer. You bring your unique business context, brand insight, and marketing intuition, and AI adds its superhuman abilities to process huge data sets, calculate forecasts and predictions, and execute decisions in real time.
So, what’s new? With Kokai, you now have even more ways to activate your own first-party data and insights as inputs to help you control exactly how the platform AI works for your campaigns. The more data and insight you input, the better our AI can understand what you value and help you make campaign decisions around those values.
Get to know our three most important AI enhancements:
1. Real-time forecasts and recommendations, right where you need them
We know you want to make the best campaign setup and optimization decisions, but you might not always have the time to pull reports and find the right data. That’s why the platform now shows you specific data, insights, and recommendations right in the context of your workflow to help you make better campaign decisions faster.
Within the platform, you will see a live snapshot of key campaign health metrics (pictured above), including your forecasted spend. As you navigate from advertiser to campaign to ad group levels, the view will change to show you metrics that support the decisions you usually make at that level. These metrics update in real time as you make changes to help you predict their impact before you set them live.
Relevance is another key metric you’ll see in the real-time snapshot. The platform uses your seed (consisting of conversion data or the closest available proxy) as the signal for which type of people are most valuable to your campaign. It then calculates relevance, which quantifies how similar your seed is to the people you’re expected to reach with your current targeting. If the platform shows a low relevance score for your campaign, you can make changes to your audience or inventory strategies and see the impact of those changes in real time.
AI within the platform will offer specific guidance and recommendations that appear alongside your setup decisions. For example, when you set your base bid and max bid, AI-powered guidance will surface to show you how other similarly targeted ad groups are bidding across the platform. These insights can help you set competitive bids from the start, without the need to guess and check.
2. New ways to control and activate your AI co-pilot
You are always in control of your campaigns, and we encourage you to use the available inputs to express your strategy and preferences in the platform. You can decide which decisions to control yourself, and which you want to delegate to your AI co-pilot.
In the Budget Allocation tile [B], you can specify budget rules or choose an automated and value-driven budget strategy. When you select the value-driven option, the platform AI can help you manage your allocations, monitor the performance and relevance of your ad groups, and automatically reallocate budgets among your ad groups to help increase value for your campaign. This is a way to simplify a key part of your campaign management workflow.
Bid factors are valuable inputs you can use to express your strategy in the platform, and we encourage you to apply your own bid factors on as many dimensions as possible. Your bid factors help our AI prioritize the kind of impressions you value most, by either increasing volume or increasing bids. You can use the Koa Optimizations [K] tile to review and decide where you want to activate AI optimizations instead. When you turn on Koa Optimizations [K] for a dimension, you can review a log of any changes the platform makes to your campaign.
When you create audiences, you can use relevance and value scores to choose from the hundreds of thousands of segments available in the Data Marketplace [Dm]. No one has time to monitor the performance and relevance of every individual segment over the flight of a campaign — but that’s a fitting job for your AI co-pilot. When you use Koa Audiences [Ka], you enable the platform to actively optimize your audience targeting by removing segments that are less relevant or that perform less well, and adding segments that can help deliver more value.
3. Value-based buying, powered by AI
The most important application of AI within programmatic advertising is for real-time decisioning. Your campaigns running on our platform evaluate up to 15 million ad opportunities per second, deciding whether to bid and if so, how much. These decisions require the platform to consider potential tradeoffs among pacing, performance, quality, and more. All at once, and in a fraction of a second.
With Kokai, our platform AI considers each impression opportunity individually and calculates its value to your specific campaign. In this context, “value” is the sum of many AI calculations that we have distributed across the platform, including KPI prediction, relevance, your bid factors, inventory quality, and more.
While the platform will always stay within the constraints of your max bid, rails, and other campaign parameters, these AI enhancements allow it to bid differently based on value. At times, AI will identify high-value impressions and bid slightly higher, and in other cases, lower when it recognizes that you can get the same or even more value at a lower bid.
Value-based buying in programmatic advertising is important because a campaign will often see thousands of impression opportunities at once that each meet the advertiser’s minimum criteria based on audience, geo, inventory, and other rails. They all qualify, and the campaign could buy any one of the impressions, but making the right choice is paramount. With Kokai, the platform effectively stack ranks qualifying impressions to help you buy the one that represents the best value for your campaign.
You’re still in control
While our platform’s AI makes calculations and decisions in mere milliseconds, you are still in the driver’s seat. Your seed will power relevance calculations throughout the platform, which is why we encourage you to use conversion data (or the closest available proxy) to create your seed during campaign setup. Your bid factors will tell the platform where to bid more or bid less, and your goals will steer the platform’s KPI predictions and optimizations.
Your AI co-pilot in The Trade Desk platform has never been more powerful, but you’re still the one in control.
Get Started with the Kokai Workflow Guide
New to The Trade Desk? Contact us to learn more about Kokai today.
Disclaimer: This information is provided solely for background and is not a representation or guarantee of any future performance.