How Samsung Ads’ CTV audience optimisation tool can ‘learn and adapt’ as you go

How Samsung Ads’ CTV audience optimisation tool can ‘learn and adapt’ as you go

Samsung Ads Europe has launched a machine-learning advertising solution that automatically optimises connected TV (CTV) campaigns.

Smart Outcomes was designed and built in-house, and uses data from more than 60m Samsung smart TVs in Europe to create advanced lookalike modelling to improve reach expansion, content promotion and customer acquisition.

The solution will “continuously learn and optimise delivery” and can be applied to advertisers’ first-party datasets through a data clean room.

Beta partner Sky Italia used Smart Outcomes to identify the most active Now app users and find lookalike audiences. Using the solution led to a 78% increase in in-app downloads and a 27% reduction cost per conversion (app download).

Advertisers can select various desired outcomes for the algorithm to optimise for, including gamer acquisition, website conversion and in-app conversion.

The tool is available in Austria, France, Germany, Italy, Spain, Switzerland and the UK for managed and programmatic campaigns. Other global Samsung markets, such as the Americas and Asia-Pacific, “support the solution”.

Analysis: Acquisition, retention, expansion

The roll-out follows the launch of Samsung Ads’ Insights Planner, which shows how many audiences a linear TV campaign did and did not reach.

According to Matt Bryan, director of analytics and insights at Samsung Ads Europe, both tools are “driven by conversations with advertisers from real problems they are telling us about”.

Where the Insights Planner addresses the challenge of understanding where viewers are in a fractured CTV world, Smart Outcomes is about “executing” on that insight.

“It helps you to better target your audience and that is driven by the largest first-party dataset in your TV viewing dataset in Europe,” Bryan told The Media Leader.

“It allows an algorithm to continually optimise for the outcome — so not only have you optimised at the beginning but, as you go through it, it continues to learn and adapt. And at the end of that, you get more effective spend and a better outcome for advertisers.”

What that means is the live campaign is “continually getting better” based on the algorithm reacting to billions of data points being fed into it, he added.

Samsung Ads launches its answer to ‘break down silos’ in TV and streaming

Bryan outlined three ways Smart Outcomes could help advertisers: first, audience targeting to acquire new customers; secondly, promoting pieces of content or genres; and thirdly, creating lookalike audiences.

He explained that while traditionally the first was the most demanded from app partners, the second element is now “really important” for retention and “keeping customers engaged”, especially as subscription VOD has started to “peak out almost”.

Finally, a lookalike audience based on “a seed audience” is crucial for clients to be able to “expand the reach” of campaigns to similar audiences outside of their first-party data.

Bryan continued: “There is so much more to do and so much further we could go.”

Media Jobs