It’s still early for machine learning in marketing and advertising. Many marketers don’t quite grasp what it is, how it works and whether it will replace them. In short, machine learning in media buying uses serial algorithms and predictive analytics to find the right audience at the right time at the best price—and no, it’s not replacing anyone.

Working within automated systems, machine learning makes quick work of countless data points to predict which groups of people will respond to an ad, then hands that information off to automated systems that place the ads. Media teams that embrace the technology can access new efficiencies, surprising insights and time better spent on bigger-picture projects.

Machine learning drives optimization with automated systems

For almost a decade, marketers have used traditional ad exchanges to bid on specific inventory against a predetermined audience. Teams have become accustomed to the hands-on nature of those mechanisms—choosing ad units, identifying prospective customers, setting price ceilings. But when it comes to evaluating performance, it is often unclear exactly what about the ads performed. Traditionally, a team would need to run dozens of experiments to find out or glean any concrete insights about how to optimize a campaign to drive more value.

Now, machine learning makes the process much easier, as it supports systems that automate the process of finding the best buy. Here, buyers set a campaign goal for the machine learning to steer toward, then set campaign parameters allowing for the widest possible data set. The machine learning models and algorithms then predict which opportunities are most valuable based on incoming data in real time as ads are served, making the adjustments needed to help ensure your campaign meets its goals at the best possible price.

To reach your goals, machines need room to work

Media planners and buyers accustomed to running campaigns on traditional exchanges may be wary of trusting machines with their work. But automated systems enhanced by machine learning work best when parameters are broad.

Teams working within these systems are advised to embrace a certain agnosticism toward placement, platforms and yes, even audience. This gives systems more opportunities to consider when assessing which will deliver the best performance. Planners who embrace the unknown may be rewarded with surprising audience insights, while media buyers can spend time on bigger-picture analysis.

Of course, optimizing for liquidity requires buyers and planners to set sound campaign goals for their campaigns. These goals are the signal that systems follow when making calculations and should align closely with real-world business outcomes. Marketers optimizing for direct-response goals will find an array of signals within the system that work well: sales, downloads, and other conversions, for example. When optimizing for brand goals, marketers may have to test which signals work best when scaling for reach, awareness or sentiment.

What it means for marketers

Boost liquidity by pairing machine learning within.
Even the sharpest media teams can’t match this dynamic duo when it comes to processing complex data sets in real time. To help ensure every dollar is being spent as efficiently as possible, take advantage of the liquidity these systems offer.
Loosen the reigns when striving for liquidity.
Broad parameters work best here. By remaining agnostic about certain elements of a buy, teams help ensure the system can access as many opportunities as possible—and find the best ones for meeting campaign goals.
Choose goals that align with real-world business outcomes.
Systems work toward these goals like a ship following a navigation signal. Choose the right one, and you can be rewarded with a more efficient campaign; choose the wrong one, and you may find yourself off course.