For an ML model to accurately predict which audience a user belongs in, it needs good data about them and it needs it while they are still on your site. Snowplow delivers real-time behavioral data and Snowplow Signals computes user profile attributes as people interact with your product, so the model works from what someone is doing now rather than what they did last week.

Conductrics learns which audiences respond to which experience, allocates each user in real time, and gives you a readable tree rather than a score.

Ezequiel Boehler, Solutions Architect at Conductrics, and Jordan Peck, Director of Field Engineering at Snowplow, will build predictive audiences from live Signals attributes and the Snowplow events you pick as the goal, then show the controls that keep the judgment with your team. What the model decides comes back into Snowplow, so you can check its work in your own data.

What you'll learn

• What a model can do with a full behavioral event stream and live attributes, rather than page views and a conversion flag
• How it finds its audiences, and what you can constrain: how many, how strong the evidence, which version runs
• What lands back in Snowplow, so the results sit alongside everything else you measure

‍

By registering, you agree that your information will be shared with our co-presentation partner, Conductrics, for the purposes of this webinar and related follow-up.