This video is part of Pega Academy course: Introduction to Pega Marketing
Identity Matching Demo
This demonstration will describe how to design strategies that recommend relevant offers by leveraging the identity matching capability.
These are the recent interactions of customer Troy on the U+ Bank's website, listed in reverse chronological order.
Troy initially browsed the website as an anonymous user. The corresponding interaction records use Troy's cookie ID (U-*) rather than his customer ID (14). When an offer is presented on the website, it is recorded as an Impression. When the offer is clicked, it is recorded as Clicked.
When Troy signed-on in the same browsing session, a correlation was established between the cookie ID (U-*) and Troy's customer ID (14). From that moment on, all previous interactions recorded for that cookie ID are considered Troy's interactions. And all future interactions by Troy in that same browsing session are recorded with Troy's customer ID.
Let's examine Troy's profile to view this combined information in a single view. This is Troy’s marketing profile from the bank's perspective. It shows all the relevant information about him. You can see a list of all interactions the bank has had with Troy under Engagement History.
This includes interactions that occurred when Troy was browsing as an anonymous user. The information about Troy’s interactions, combined with his known customer data and calculated scores, results in the Next-Best-Action recommendation. At the moment, if an interaction takes place with Troy, the Next-Best-Action is to make an offer.
Further below you can see the three most relevant offers for Troy, with an indication of the likelihood of him accepting each one.
Let's now examine how the strategy leverages identity matching to choose these best offers for Troy. The strategy that recommends offers to the U+ Bank website is defined at the Sales level in the Next-Best-Action hierarchy.
This is the strategy that arbitrates across all available sales offers and selects only the most applicable one for the current interaction.
Each of the components on the left-hand side represents a group-level strategy, which contains the decisions that select the best offers within that group.
Let's examine one of them.
The Credit Cards group-level strategy arbitrates between the available credit card offers and calculates a priority score for each of them. The formula for calculating the priority consists of three parts.
The numeric values from each of the parts are added to calculate the priority score for each offer. The first part balances customer needs with business goals. Customer needs are represented by the propensity score, which is the likelihood the customer will accept the offer. The expected revenue represents the business goals. These two values are multiplied.
In the second part, the priority is increased with 10 points if the customer has recently shown interest in the offer. This is where the automatic identity matching feature comes into play. When the customer logs in, all previous behavior from that browsing session are attributed to the logged-in customer. This increases the priority of credit card offers over the others.
In the third part, the priority is increased further if the customer is currently browsing the credit cards section of the website. A similar approach is used to prioritize the offers in other groups, all of which will be propagated to the higher level Sales strategy. The Sales strategy then selects the offer with the highest priority across all groups.
To summarize, the identity matching feature automatically associates all previous behavior in a user’s browsing session with a known customer once the user logs in. This information is combined in the strategy with other known customer information, propensity scores and business priorities to identify the best offer for the customer’s current situation.