A better approach to audience selection
We've all seen it before. A brand's unique audience persona - detailed in a nuanced audience brief - unjustly minimized into a handful of 3rd party audiences.
For example, take this food brand's persona:
Adventurous and thoughtful food consumers, who like to make choices good for the environment and their bodies, while also prioritizing getting the most out of life. They are willing to sacrifice a little taste for a healthier alternative. They are typically in a household without kids and have the disposable income to spend on unique food experiences.
This might get manually translated into three audiences: 'Foodies', 'Healthy eaters', 'Household income $150K+'. While this selection is acceptable, with the advancements in AI and LLMs, there has to be a better way.
With Dstillery incorporating OpenAI's technology that powers ChatGPT intoAudience Explorer, we finally have a way to understand and quantify the relationship between the audience persona and specific audience segments like never before.
Feels like magic
When AI just works, it feels like magic. And when you use our new LLM-based search feature in Audience Explorer, you might just feel this way, too. So how does it work?
Getting a bit into the technical details, Dstillery leverages OpenAI's embeddings to create a method for establishing a meaningful relationship between an audience brief and Dstillery's audience taxonomy.
First, Audience Explorer takes an audience brief as an input, which typically includes descriptions of interests, demographics, affinities, and other relevant information. OpenAI's embedding technology is then employed to convert this input into high-dimensional vector representations - a fancy way to say number sequences that are easy for computers to understand.
Next, these embeddings are compared with Dstillery's entire audience taxonomy, which has already been translated into vector representations similar to the audience brief.
By analyzing the vector similarities, we can identify the audience segments with the closest similarities to the text in the audience brief. This approach not only saves you time from manually selecting audiences, but also ensures that your campaigns are using the audiences most aligned with your audience persona ultimately leading to a highly targeted, more effective ad campaign.
Audience Explorer, powered by OpenAI, now enables advertisers to finally bridge the gap between your brand's unique audience personas and targetable audience segments. With Dstillery's audiences found in every major DSP, and made available as user-based segments, PMPs, and custom bidding algorithms, activating any of these audiences is just a few clicks away.