Predictive analytics is defined as a set of methods that make predictions about future events or the behaviours of entities which could for example be customers. Such a method is used to construct a predictive model from a set of historical data.
Predictive analytics finds application in an increasing number of services and industries. A traditional application is credit scoring in the banking industry. A bank would use a predictive model to define the probability a customer will default on their mortgage. The predictive model is built from the bank's customer data and could include variables such as average income, previous loans, job category etc. Services or industries and examples of applications where predictive analytics is used include but not limited to:
The ability to predict future events is a very valuable tool for any business and allows for many sorts of optimisation. A business using predictive analytics can be proactive in its business decisions rather than reactive. Knowing which customers will churn allows a telecom provider to target its marketing activities in an efficient manner, being able to identify fraudulent claims is an important tool for any insurance company, knowing which customers are most likely to respond enables the optimisation of a direct marketing campaign.
An appropriate metaphor is to consider predictive analytics as a "fuzzy crystal-ball", fuzzy because it only gives the probability of an event. How could your business benefit from a crystal-ball, even if its not perfect?
In short Ofvigo's predictive modelling service is automated robust and fast and does not require expertise in building predictive models. The Ofvigo predictive modelling service uses Ofvigo's own proprietary predictive modelling algorithm. This algorithm has a very important characteristic - it is very robust and makes few assumptions about the data used. Contrast this with the widely used logistic-regression algorithm for building predictive models where considerable expertise and time is required to ensure that the data conforms with the assumptions inherent in logistic regression and to ensure stability of the model. The robustness of Ofvigo's approach means the predictive model building process can be automated and accessed as a service.
The recommendation platform provides an easy to use front-end for an ecommerce business who would like to increase their cross-sell or click-through rates by using intelligent recommendations. Ofvigo's intelligent recommendations are generated using the generic predictive modelling service. However for the recommendation platform there is no need to know anything about predictive modelling. The platform automatically collects the required click data from the ecommerce site, decides when to build predictive models and then generates personalised recommendations for each customer from the the models. The whole process is completely transparent to the ecommerce site administrator. Now you can benefit from predictive analytics without being an expert.