Propensity modelling is a statistical approach and a set of techniques which attempts to estimate the likelihood of subjects performing certain types of behaviour (e.g., the purchase of a product) by accounting for independent and confounding variables that affect such behaviour. Normally, this likelihood is estimated as a probability known in propensity models as propensity score. This functionality makes propensity modelling a popular technique in various fields including economics, business, education, healthcare, marketing, and more.
Why Propensity Modelling?
Why should we use propensity modelling to infer the causes and implications of behaviours if we have other alternatives such as conventional randomized trials or A/B testing? The thing is that we can’t always rely on these statistical methods in the real world. There might be several scenarios where real experiments are not possible:
- Sometimes management may be unwilling to risk short-term revenue losses by assigning sales to random customers.
- A sales team earning commission-based bonuses may rebel against the randomization of leads.
- Real-world experiments may be impractical and costly in certain cases when the same data or participants can be modelled through quasi-experimental procedures or when historical data is enough to produce actionable insights.
- Real-world experiments may involve ethical or health issues, for example, studying the effect of certain chemicals.
Propensity models are useful in the following situations:
Test marketing campaigns: With the propensity model, we can test marketing campaigns in the shopping locations and even cities before the fact to assess how effective the campaigns would be given a set of parameters similar to other shops or cities where our campaigns have been run already.
Measure marketing efficiency: Propensity models can analyse marketing effectiveness by channel to reveal various trade-offs between investments in different media and improve marketing ROI estimates.
Reduce costs: Propensity score can increase the ROI of your marketing campaigns if several simple rules are followed. First, contact people, for whom a campaign has a higher chance of success and, therefore, is likely to result in increased revenue from the campaign. Secondly, understanding propensity scores by customer groups you can give each customer the minimal offer needed to attract them. Such offers will be more advantageous when used with customers with a medium propensity than for those with the high propensity, as these reduced offers will lead to significant savings.
Prevent fraud: Propensity scoring can determine if a customer applying for your services is a bad debt customer or a fraudster. A propensity model embedded into an application can determine how to handle the customer – completely deny the service or offer an alternative product / package. This type of fraud prevention results in the maximization of revenue and minimization of risk.
Do you need help? Do you want to develop customer propensity model for your business?
Contact us
Unit No. 450, Mastermind-1 Premises, Royal Palms, Goregaon (E.), Mumbai – 400065
