According to the IT consulting firm Gartner, Business Intelligence is a generic term that includes applications, infrastructures, tools and best practices to enable access to information analysis to improve and optimize decision making.
BI helps us find the answers to previously known or established questions.
Beyond this definition, BI is fed by mainly structured data sources, such as statistical data, external tables, ERP and CRM systems and structured information of activity generated in web pages, among others, to help us find the answers to previously known or established questions such as
- What type of insurance generates the highest profits?
- What policies are taken out by customer segments with common characteristics?
- What type of insurance is taken out on certain dates or at certain times of the year?
- What are the characteristics of my top 10% of customers?
- What promotions are most attractive to my customers?
- What policies are taken out by the same segment of customers?
- How are my results going against the budget?
- What kind of expenses are increasing?
- Where is 80% of my sales, costs and profitability concentrated?
- What objectives are being met?
But on many occasions the results are difficult to understand and we find it hard to understand what circumstances have led to a certain result; our traditional thinking would lead us to try to find out the roots of what causes certain events, to understand why.
Big Data offers an alternative that in many occasions will be of more value than the effort to find out the meaning of things, it presents us with results in a fast way when operating with the totality of the data. The conclusions are close to reality and they present us with the "what" of our environment, that is, we obtain information about "what" is happening in the world.
Big Data also helps us to find those issues that you can't find or haven't been raised.
It is therefore, that it offers us new strategies within the insurance sector as:
- Fraud detection algorithms. Helps insurance companies reduce the number of fraud reports and avoid large losses by applying analytical models to build a risk profile associated with an identity and creating a risk-based score that predicts the likelihood of fraud on an application.
- Costs and outsourcing of litigation With Big Data, insurance companies can design more complete business rules based on a strict scientific method to assign a litigation to an external legal team or to company lawyers to improve the optimization of the use of resources.
- Marketing Mix. The aim is to develop an optimal marketing model where a balance between marketing initiatives is achieved, maximizing the return on investment. The objective is to predict who will respond to the offer of a product or service. To determine the Marketing Mix, the analysis to be carried out can be based on the past behaviour of a similar population, on the response obtained to an equivalent strategy, or on some logical element susceptible to statistical analysis
- Renewals, propensity to abandonment and retention of clients. Due to the high level of leakage in the sector, an insurance company needs to optimize its collection processes through segmentation based on the propensity to default. Big Data's advanced analytics techniques allow the identification of customers with a higher propensity to abandon and group them in categories to design specific retention strategies for each risk segment.
- Campaign management. Business Intelligence is not enough to meet the new needs, these traditional systems are limited to the past and the vision that insurers require today must also include the future. BI must be complemented with the incorporation of predictive models that allow to gain in understanding. The Big Data solution chosen must allow a segmentation of the portfolio using a scientific method that identifies the ideal channels through which to influence each customer or profile.
After this brief summary of the characteristics of Business Intelligence and Big Data, we can clearly conclude that Big Data does not replace BI, but complements it by providing prediction.
