We are in 2022! It’s going to be a great year. The Vidora team is truly grateful and excited to be working with such a great set of customers and partners.
One of the key questions we ask ourselves is how can Vidora provide the most value to our customers? To answer this question, our team does a careful analysis of both our customer needs and broader market trends. There are a couple broader trends we are seeing and focusing on in 2022: (a) an increasing need for real-time inference to engage with anonymous and first-time users in particular and (b) businesses transitioning from passive machine learning “predictions” to more proactive machine learning “decisions”.
Real-time inference enables a business to make a fast decision (<100ms) based on how the user just engaged. The performance increases combining batch with real-time features can be substantive. Customers can see upwards of 200%+ increase in conversions when adding real-time features.
Real-time also plays into broader industry trends. As businesses try to engage with anonymous and 1st time users, real-time becomes a critical tool. Vidora’s framework has a unique implementation of real-time which combines both in-session and historical features in the real-time inference. The use of complex batch features combined with up-to-date behavioral signals results in the best possible performance.
Vidora will continue to expand our real-time Decisioning SDK (read more here) and Cortex functionality to empower our customers with the right tools and frameworks to leverage real-time decisions across their business.
Machine Learning Decisioning
Machine learning is often thought of as making “predictions” about the future. However, predictions are inherently a passive analysis of user behavior. For instance “predicting who will churn” or “predicting who will convert” provides information on the user, but doesn’t tell a business how to impact that user’s behavior in a positive manner. What typically drives the most value for businesses are active decisions about what action to take on a user : “how do I engage a user to reduce churn” or “what message do I use to increase conversions”. Predictions are a passive analysis of what a user will do, while decisions are a proactive engagement to increase the user metrics you care about.
Enabling businesses with machine learning decisioning, requires new tools from both a modeling perspective (i.e. prescriptive models like Uplift) and machine learning operations perspective (i.e. weighting decisions outcomes, monitoring decisions). The Vidora team is already busy building out more tools and features to empower our customers to leverage machine learning decisioning across their business.
2022 and Beyond
We are incredibly excited about 2022. We look forward to seeing how our partners leverage machine learning to drive dramatic value through automation and optimization. Thank you all for being part of this incredibly exciting journey! Please reach out to email@example.com anytime if you have thoughts, questions, or want to discuss real-time machine learning decisioning.