Since Beeswax introduced our Bidder-as-a-Service™ platform in 2015, we’ve seen a wide variety of marketers, publishers, and adtech companies jump on board with the goal of innovating and owning their own tech stack for programmatic buying. We’ve seen our customers innovate around workflow, first and third-party data, and attribution. But the most prevalent use case for our solution is custom optimization, as we allow sophisticated buyers to write and control their own bidding algorithm. It’s something we call “Bring Your Own Algorithm” or “BYOA”, and it’s made possible through our Bidding Agent product.
Today, we’re excited to announce the next stage in empowering buyers to BYOA: Bid Models. Bid Models allows customers to build, test, and execute BYOA programs quickly and easily, with minimal technical overhead while still having the power to control bidding using any combination of multivariate inputs.
Check out our webinar to understand why created Bid Models and how it works:
Beeswax now has the widest and most comprehensive solutions for BYOA:
- Standard Strategies enable built-in algos for simple outcomes
- Bid Modifiers empower ad ops and analysts to quickly and easily modify bidding behavior based on auction variables or user scoring
- Bid Models allow analysts or data scientists to build comprehensive multivariate models and upload directly to the Beeswax cloud
- Bidding Agents are the industry’s only solution for sophisticated customers to deploy their own code in the cloud for a completely bespoke bidding environment.
While we’re on the topic of BYOA, let’s address a key question we are sometimes asked: Why do I need my own algorithm? Good question, thanks for asking. It’s true, the big, integrated platforms have more data to tune their models across thousands of customers. But, have you ever thought about how your data is also informing those models, making them perform better for your competitors? In fact, our customers find that when they tune their models to their specific KPIs they can get better results, cheaper, than through a generic, non-transparent model run across advertisers.
To get started using Bid Models, here’s some documentation:
- Overview: how to create a Bid Model
- Technical tutorial: GitHub
- Manifest of available fields to use in a Bid Model: Google Doc
- Sample Bid Model files: GitHub
Ari is a recognized product leader in ad tech and SaaS. He served as a VP at DoubleClick and a Director of Product Management at Google, where he led the buy-side product suite. More recently he was the SVP, Product Management at Appnexus and the EVP, Product Management at publicly-traded Bazaarvoice. He is the creator of the “VAST” standard for video ad serving, a patent holder on Nielsen’s OCR (online GRP), and contributor to AdAge, AdExchanger, and other publications.