Role
Optimisation Lead
Company
HSBC (APAC)
Platform
Website
Nature
Proof of Concept, Marketing Technology, Development
Dynamic Message
The Project
Traditionally users were receiving generic, non-personalised content that didn't reflect their banking needs or interests. The lack of real-time personalisation meant missed opportunities for relevant product recommendations and reduced engagement with promotional content.
The initial idea was to display different sets of messages, in hero banners, side promotions, and tile messages, to specific personas defined in Adobe Audience Manager (AAM). I took this as a challenge and created a proof-of-concept (POC), which received an overwhelming response from the business stakeholders.
After opening the door to new possibilities, I kept enhancing the implementation, making it more flexible and scalable. The solution was gradually rolled out to other Asia regions, then EMEA and the Americas, and eventually became a global solution within HSBC. This project was one of my most iconic and successful initiatives at HSBC.
Hyper-personalisation
The dynamic messaging solution was a resounding success, with over 50 personas being created and running globally across HSBC's digital platforms
In the second phase of the project, I took the initiative to enhance the solution further by transitioning from using Adobe Audience Manager to leveraging profile scripts in Adobe Target. This allowed for real-time classification and customisation of the personalisation algorithm
Using propensity scores to define personas based on the frequency and recency of customer visits, and apply it to logout page and individual product pages
Unlimited Capabilities
The solution of the dynamic messaging and its evolution demonstrated my ability to leverage existing technology stack, continuously innovate and drive the project towards deeper levels of personalisation and customer-centricity.
What makes it even more interesting is, we can develop further A/B tests for each persona, so that we can evaluate the performance of different messages. Moreover, it can also combine with machine learning algorithms to optimise the classification of personas.
