One of the issues that’s eluded most retailers is the conversion of casual browsers into purchasers. Even though more consumers are online today, the average conversion rates range from 2%-4%. Most marketers turn to offers and promotions to solve this problem. But offering the same promotions sitewide erodes margins for all visitors. The retailer wanted to determine which visitors to target for offers and promotions.
Here are the challenges the retailer faced:
- How to identify visitors who might be interested in buying but are hesitating for various reasons? In other words, they are “on-the-fence” and can be influenced.
- How to engage the on-the-fence visitors with an incentive or information that is right for them but also does not erode business margins?
- To take these actions before the visitor leaves their site.
To better understand the right visitors to target, the retailer partnered with Session AI. In particular, the retailer leveraged Session AI’s RevPredict dashboard. Powered by its in-session Early Purchase Prediction (EPP) model, the dashboard reveals the predicted purchase outcomes early on in a visitor’s session. The platform then uses these predictive insights to determine the best action to engage each visitor in real-time.
The retailer used Session AI’s in-session marketing tools to group its site visitors into:
- Those who will not buy — no point in showing them an offer.
- Those who will buy — the best strategy is to not interfere in their journey.
- Those who are undecided about a purchase, or are on-the-fence — the retailer focused on this third group of visitors
By leveraging in-session intent prediction, the Retailer was able to identify influenceable visitors and send personalized, real-time engagement for them. These visitors are revenue opportunities that would have been lost had the retailer not proactively acted to nudge them in their customer journey at the right time. As a result, the retailer achieved a 20% average lift in revenue per visitor.
How did it happen?
1. Form an early purchase prediction (EPP):
The Retailer deployed in-session marketing to readily identify influenceable visitors. It then activated personalized interactions to motivate them to make a purchase while they were still on the site. The in-session intent analysis took into account multiple factors, including the visitor’s past purchase patterns, preferences, number of times they have viewed a product in the past, as well as their current behavior on the site. Based on these and other information, the EPP model inside the platform calculated a visitor’s purchase propensity and identified those within an influenceable range.
2. Incorporate environmental data:
After identifying on-the-fence visitors, the EPP models quickly determined the best action to show each visitor after considering their historical purchase decisions and preferences. For example, is the visitor price conscious? Do they need social proof for product confidence? Is convenience the most important factor for them? Additionally, the platform considered data such as the location, time, in-store inventory, or any external events that might impact the purchase decision.
3. Deliver real-time actions:
At this stage, the retailer had the opportunity to influence the buying behavior across a range of possible personalized engagement such as the following:
In summary, the U.S. retailer was able to create differentiated experiences for every customer’s unique needs, particularly those who would have bounced off the website without making a purchase. The retailer achieved a 20% average lift in revenue per visitor by offering them a variety of engagements that are timely, helpful, and relevant.