Financial Services
Next-best-action engine
A 4% increase in trading volume, delivered by connecting client behaviour data to marketing automation inside a wealth management app.
The client was a global wealth management organisation serving high-net-worth and mass-affluent clients through a proprietary digital platform. Clients used the app to track portfolio performance, market movements and account activity. The app was their primary digital relationship with the firm.
This engagement was delivered prior to the founding of LoweLabs.ai. It's included here because it shaped the firm's approach to connecting analytics, marketing technology and commercial outcomes. It is the same discipline LoweLabs now applies to client engagements.
The firm had already licensed a white-labelled SaaS product for web-based analytics, instrumented against the client-facing app and its supporting APIs. It could tell the business what clients were doing inside the app: what they viewed, how often, and for how long.
What it couldn't do on its own was act on that information. Analytics without an activation layer is a report, not a capability. The firm needed a way to turn behavioural data into segmentation, personalisation and next-best-offer decisions, delivered back to the client inside the app, without building a bespoke engagement engine from scratch.
Two constraints shaped the work. First, the analytics data had to be trustworthy before anything downstream could be trusted, which meant the app's existing instrumentation needed scrutiny, not assumption. Second, the business case for the investment depended on proving a commercial link between personalisation and revenue-generating behaviour, not just engagement metrics.
The engagement was structured around closing the gap between "we can see behaviour" and "we can act on it profitably":
Instrumentation audit and correction
Reviewed and rebuilt event tracking across the app to ensure the analytics platform was capturing genuine user behaviour, not artefacts of poor tagging or incomplete event definitions. This was treated as foundational: every downstream decision was only as good as this data.
Integration with IBM Unica
Connected the SaaS analytics output to IBM Unica's marketing automation engine, turning passive behavioural data into an input for active decisioning.
Segmentation and next-best-offer logic
Used the integrated data to build client segments and determine the next best offer, advertisement, or piece of information to surface to each client, delivered natively within the app experience.
Attribution modelling
Built a model connecting exposure to personalised content and offers back to actual trading activity, closing the loop between marketing action and commercial outcome.
The client app and its supporting APIs fed usage events into the white-labelled analytics platform. Once instrumentation was corrected, this data was passed into IBM Unica, where it drove segmentation and next-best-offer logic. Unica's decisions were then served back into the app as personalised content, offers or information, closing the loop from data capture to client-facing action. A parallel attribution layer tracked trading activity against offer exposure, allowing the business to measure whether personalisation was actually moving the metric that mattered.
The programme delivered a measured 4% increase in trading volume attributable to personalised, next-best-offer content, meeting the threshold set in the original business case for the investment.
The application itself was recognised as a leading digital tool in wealth management at the time, validation that the combination of accurate instrumentation, activated analytics and measurable attribution had produced a genuinely differentiated client experience, not just a technically impressive one.
Instrumentation is the foundation, not a formality.
Segmentation and personalisation logic is only as reliable as the event data feeding it, and that data is rarely trustworthy by default. It has to be audited and corrected before it's used for anything commercially significant.
The value sits in the integration, not in either tool alone.
Analytics platforms show behaviour; marketing automation platforms act on inputs. Neither delivers commercial value in isolation. The return comes from the connection between them.
Attribution is what makes personalisation defensible.
Engagement and click-through metrics are easy to generate and easy to argue about. Tying personalisation directly to a commercial metric, in this case trading volume, was what protected the investment case and justified continued funding.
Next steps
Be Relevant.
If your organisation hasn't captured the value of personalisation, or having trouble turning their data into demonstrated economic outcomes, let's talk about closing that gap.