Predictive models use the latest advances in machine learning algorithms to predict an outcome – whether at an individual level such as how likely someone is to buy a product, or a macro level such as likely sales given investment across marketing channels.

Our models guide better business decisions by incorporating a variety of different data sources, from surveys and databases to social and API sourced data.

All of our predictive modelling work serves to optimise marketing investment decisions. We can apply predictive models to address broad marketing questions (for example, to understand the effectiveness of different marketing activity) and to address specific marketing challenges (for example, to identify individual customers who are at most risk of switching to a competitor).


Our Predictive Models Offer

Churn Risk

Acquiring a new customer is anywhere from 5 to 25 times more expensive than retaining an existing one. So, for any business, keeping hold of the right customers is critical. Our Churn Risk models identify customers most at risk of leaving, allowing pre-emptive, targeted intervention.

We use the latest machine learning algorithms and Natural Language Processing to pinpoint at risk customers. Our models are strengthened by blending a variety of historical data, including customer behaviours, customer touchpoints, customer logs, Net Promoter Score® surveys and more.

We can produce a snapshot of churn risk by customer, or embed advanced models in your systems so you can score your customers over time.

Propensity Models

Successful marketing reaches the right customers with the right offer through the right channel at the right moment.

Propensity models predict consumer behaviour, thereby enabling marketers to steer resource allocation and highlight which customers warrant specific attention. For example, you could reduce direct mail spend by targeting only customers who are most likely to respond. Or identify customers who are more inclined to complain.

We build propensity models for your current customers today, and your prospective customers of tomorrow.

Recommendation & Personalisation

Recommendation engines are ubiquitous across the eCommerce landscape, but a small number of transactions is often not enough to understand what an individual may do next.

Our learning algorithms help you stay one step ahead of the consumer at all times, and increase your cross-selling conversion.

We achieve this by combining the transactional data with as much non-transactional data as possible, such as website engagement.