How a leading retail bank achieved real-time 1-to-1 personalization

This is the first of a series of blogs I’m writing to publicize a new study we have commissioned Forrester Consulting to undertake. Our December 2018 Total Economic Impact of Celebrus study is a Forrester return on investment investigation which has been combined with a survey designed to reveal activities, challenges and opportunities in the marketplace which relate to Artificial Intelligence (AI) and Machine Learning (ML) technologies.

I’ll focus on the case study element for this blog. Celebrus wanted to prove to the financial services world that our solution can transform the effectiveness of AI and ML to achieve genuine 1-to-1 personalization. We had the confidence to subject our client’s implementation of Celebrus to the rigorous tests laid down by Forrester, to demonstrate that Celebrus delivers genuine and quantifiable benefits to retail banks. This warts-n-all approach gave Forrester unprecedented access to our client in the absence of any of the Celebrus team, as well as the freedom to scrutinize and verify all financial data.

Our client is a leading northern European bank, who some time ago decided to embark on a journey of digital transformation. Specifically, they made the commitment to become a data-driven organization, seeking to leverage customer data to drive revenue and enhance the customer experience. The bank decided to implement an Artificial Intelligence solution to generate next-best-actions which could personalize the content that each customer is presented with when engaging with the bank’s many digital channels. During the preparatory work which they undertook to scope the project and to define best practice, the bank quickly reached the conclusion that the success of this project would depend on the quality of the customer data flowing into their chosen decisioning solution.

Quite simply, this bank recognized that high quality data is needed to enable accurate next-best-actions, which in turn result in every individual customer being presented with highly relevant web or app content and the most appropriate product promotions. As with all of their technology acquisition decisions, this enterprise bank insisted that prospective suppliers demonstrate their capability and success in previous projects. When it came to selecting a solution capable of capturing the most granular and accurate customer interaction data possible from the bank’s digital channels and connecting it in real-time to their enterprise AI solution, the bank chose Celebrus due to our large number of highly successful implementations in leading global banks.

To provide a clearer understanding of how all of the systems involved work together, let me talk you through the bank’s 1-to-1 personalization project step by step.

• It all starts with the customer interacting with the bank’s digital channels - it’s website, mobile app or perhaps its ATMs. The bank has decided it wants to capture this interaction data to make use of it for personalization, so it deploys Celebrus for this task. The Celebrus real-time CDP identifies logged in customers and captures all of their experience data in the simplest way possible, eliminating the need for tagging
• Now that this data has been captured, many CDPs would simply connect all of the data to the third-party application that had been chosen to receive it, but Celebrus is able to extract more value from this data. Celebrus can detect and identify signals within customer behavior that indicate either opportunity or threat. This could be sustained browsing for a particular product or suspicious behavior which may indicate fraud. Celebrus can be configured just to connect these signals to the bank’s chosen AI solution to enable the third-party solution to operate more efficiently
• Celebrus then connects these signals to the bank’s AI solution in genuine real-time, meaning that data is received by the third-party application within 200 milliseconds. This capability is unique to Celebrus for two reasons: firstly, other CDPs are unable to connect with best-of-breed enterprise solutions, preferring to build out large numbers of integrations with mid-market applications; and secondly, despite their ‘real-time’ claims, no other CDP is able to connect data anything like as quickly as Celebrus
• The bank’s AI solution then receives the Celebrus behavioral signals and generates a next-best-action, which determines the content that should be served to the customer
• The bank’s AI solution connects with their content management system to serve this content to the customer, who receives an entirely bespoke browsing experience determined by his behavior and the details within his customer profile

The entire process above, from customer interaction to bespoke content being served, takes a total of 500 milliseconds or less – anything longer and real-time personalization would not be possible because the page would have loaded before the next-best-action derived content had arrived.

And the main reason why we commissioned Forrester to undertake this Total Economic Impact study and to expose the financial realities of this project?, It’s because we are well aware of how beneficial and profitable real-time personalization really can be for banks. Indeed, this survey revealed that our banking client is on course to generate profits of $5.4m in the first five years.