The current solution to determining the engagement of a customer takes a blanket, narrowly focused approach. Most simply it looks only at when a customer has churned and derives this from a one-size-fits-all metric such as; ‘has not purchased in the last 2 months’.
Although this approach is scalable it fails to recognise the individual buying habits of the customer. Those that buy frequently should be defined as disengaged after a much shorter zero-purchasing period compared to those that buy more infrequently.
We modelled the behaviour of each individual customer so as to detect changes in their specific behavioural patterns. By spotting these changes on a weekly basis we were able to send out targeted marketing campaigns months before the current solution.
Another advantage of building customer-specific models is that we are able to identify those customers who show an increase in their level of engagement. By identifying winners, the client was able to optimally allocate resources to nurture that relationship and drive further revenues.
Lastly our use of inferential statistics, a computationally efficient modelling approach, gave us the capacity to run the analysis across tens of thousands of customers. The application of change point analysis further enhanced our ability to scale, providing the capability to automate the identification of behavioural changes across the entire customer base.
the number of days earlier that marketing campaigns can be triggered
hours saved by automating the detection of a customer's behaviour change
minutes to run the analysis
55% more reliable bug identification
We identified errors in C and C++ programs faster and more accurately than the existing technology.Learn more
1000X faster cable network planning
We created a way to plan fibre round a town in seconds – more than a thousand times faster than the existing approach.Learn more
Mobile cells optimally placed
We placed mobile network infrastructure to maximise coverage whilst minimising equipment costs.Learn more
Mariflor Vega Carrasco
Thank you for contacting us, we will come back to you as soon as possible