Understanding Anonymous Visitors Using In-Session Intelligence

As marketers, we’re being asked to prepare for a world without third-party cookies, where retargeting some consumers after they leave our site will be much less reliable, if not impossible. This challenge is particularly significant for anonymous site visitors who haven’t registered, nor shared any personal contact information. Meanwhile, even for known visitors on whom PII data is available, increasing consumer data privacy regulations continue to further restrict how that data may be used in core marketing practices to acquire, convert and retain customers.

What if there was a way you could understand each visitor on your site without using any PII data? From just their first few clicks. While they are still on your site. Whether they’re known or anonymous. First-time or repeat.

This new type of intelligence becomes possible when we focus on what’s happening right now in the session to make predictions. Readiness-to-buy in the current visit, or the detection of first signs of friction and abandonment, are a couple of examples of this type of in-session intelligence. Let’s take a deeper look at the first example.

Leading brands are already using machine learning models to reliably predict readiness-to-purchase using just the first five clicks from every visitor’s session data. This provides an “early purchase prediction” score that can be used to hypertailor one-to-one actions without the need for any PII data. Here are some ways to help you start thinking about how to use such in-session intelligence.

Convert Anonymous Visitors Before They Leave

Retailers can use the early purchase prediction (EPP) score to identify anonymous visitors who will not buy under normal circumstances but can be influenced. These on-the-fence shoppers can then be given a limited-time offer (e.g., valid for the next 30 minutes to 60 minutes) to influence them to buy in the same visit and secure immediate business gains. Marketers can now be far more strategic and effective compared to simply splashing a 10 percent off sitewide offer in exchange for a personal email data grab.

Save Margins by Suppressing Offers to Highly Likely Buyers

The EPP score can also be used to identify visitors who are on the site and are highly likely to buy as is — so-called “sure things.” By pulling back sitewide offers and instead using personalized offers based on each visitor’s EPP score, retailers can suppress offers for “sure things” and drive margin savings in the process. Alternatively, retailers could reallocate the offers budget — from “sure things” to “fence-sitters” — and use variable offers to maximize conversions.

Leverage In-Session Scores Through Triggers

In-session propensity scores could also be used in other marketing campaigns, such as to activate highly likely buyers who didn’t purchase with timely email reminders of changes in product prices and availability. In-session machine learning scores can be used to identify a range of such time-sensitive “triggers” which will provide marketers new opportunities to enhance user experience, increasing both customer satisfaction and conversion rate.

Leading brands are already experimenting with sophisticated in-session ML models. Early results are showing impressive results, not only in terms of revenue lift and conversions but also in terms of brand trust and customer experience. In-session intelligence will play a pivotal role in enabling tomorrow’s privacy-first marketing, where experiences are based on real-time context and all visitors are equally understood.

Manish Malhotra is the co-founder and chief product officer at ZineOne, the industry’s first in-session marketing platform that helps brands deeply understand anonymous site visitors.