Predictive analytics is an area of statistical analysis that deals with extracting information from data and using it to anticipate future trends and behavior patterns based on a variety of techniques from statistics and data mining. The core lies in capturing relationships between explanatory variables and the predicted variables from the past occurrences and exploiting this information to predict future outcomes.
In business, predictive models exploit patterns found in historical and transactional data to identify risk and opportunities. Models capture relationships among many factors to allow assessment of risk or potential associated with a particular set of conditions, guiding decision-making for candidate transactions. Predictive analytics are also applied in insurance, telecommunications, retail, travel, health care, pharmaceuticals and other industries.
How It Works
Unlike the standard business reporting and sales forecasting methods, predictive analytics offers actionable projections for each customer. This special form of business modeling foresees each customer purchase, response or cancellation, predicting the individual behavior of existing or prospective customers under certain conditions. Naturally, per-customer predictions are a key to allocating marketing and sales resources. For example, by anticipating which product features each customer will respond to, you can appropriately target the right segment of customers.
The central element of predictive analytics is the predictor, a variable that can be measured for an individual or other entity to foresee future behavior. For example, an insurance company is likely to take into account potential driving safety predictors such as age, gender and driving record when issuing car insurance policies.
Multiple predictors are combined into a predictive model which, when subjected to analysis, can be used to forecast future probabilities with an acceptable level of reliability. In predictive modeling, data is collected, a statistical model is formulated, projections are made and the model is validated (or revised) as additional data becomes available.
Predictive Models in Practice
The real trick is to find the best predictive model. There are many kinds of models, such as linear formulas and business rules. And, for each kind, there are all the weights, rules or other mechanics that determine precisely how the predictors are combined.
PAN style="mso-bookmark: _Toc192401021">
Data mining is a component of predictive analytics that entails analysis of data to identify trends, patterns or relationships among the data. This information can then be used to develop a predictive model. Predictive analytics, most predictive models and most data mining techniques rely on sophisticated statistical methods, including multivariate analysis techniques such as advanced regression or time series models.
There are three steps where business expertise is needed to direct predictive analytics:
- Defining your prediction goal;
- Evaluating the prediction results and redirecting; and
- Deploying your prediction model.
Increasingly, organizations in virtually every industry around the world are realizing the benefits of using data to align their current actions with their future objectives. By incorporating predictive analytics into their operations, companies have gained control over the decisions they make every day so that they can successfully meet their business goals. Predictive analytics is increasingly used in retail, telcom, insurance, financial services and the pharmaceutical industry. The most successful and rewarding predictive analytics processes:
- Enhance customer retention and loyalty,
- Identify cross-sell and up-sell opportunities,
- Identify customer lifetime value,
- Include fraud detection,
- Portfolio analysis and
- Direct marketing.
With the advancement of technology, better services and competitive pricing, telco companies face increasing competition for customers. Predictive analytics helps companies identify at-risk and most valued customers to develop customer retention strategies, thereby increasing customer retention, acquiring profitable customers and creating effective cross-sell and up-sell strategies.
With the increasing importance of customer retention, loyalty and churn in the context of customer lifetime value, companies are keenly following customer churn and the scale of efforts required for an appropriate retention campaign. In the diversity of consumer business today, the mass marketing approach cannot succeed; it is necessary to undertake customer value analysis along with customer churn predictions to make marketing programs target more specific groups of customers.
Predictive modeling helps identify subsegments of the customer base that are likely to churn and provides a well-identified segment to target with retention programs. Accurate predictive models can:
- Predict churn propensity into a subscriber base,
- Identify churn’s key indicators and
- Effectively target subscribers with proactive retention campaigns prior to churn.
Investment on attrition models and retention programs can provide higher ROI than spent on fresh acquisition because attrition models can save millions from lost revenues.
Another area of predictive analytics in telco is to identify prospective customers for effective cross-sell and up-sell opportunities, thus uncovering a profile of the customers who purchase numerous products or upgrades.
Study of usage patterns, payment profiles and demographic credentials are among other factors to be considered for selling additional products and services. The key benefits of applying predictive analytics for cross-sell and up-sell marketing strategies are:
- An identified set of customers with high probability of buying specific products as a key target for cross-sell and up-sell campaigns;
- Increased average revenue per user by offering best-fit products to customer; and
- Crucial information on ideal product/services bundles for each customer profile.
Successful insurers credit predictive modeling for improving identification and segmentation of insurance risks, leading to better underwriting, pricing and marketing decisions and smarter management of the insurance business.
What can predictive modeling do for the insurance industry? It can help insurers improve their rating plans by identifying mispriced risks. By analyzing distributional relationships in insurance databases in a multivariate framework, predictive modeling identifies assumptions that give misleading results. For example, when considering the relationship between insurance losses and age, and between insurance losses and prior accidents, younger drivers and those with prior accidents respectively cost more to insure.
Factors influencing customer attrition is another area where predictive analytics is used to estimate and devise appropriate retention strategies for a high value customer likely to lapse.
Predictive analytics in the banking industry is a dimension of BI that allows the assessment of risk and opportunities. In retail banking, this process translates into questions such as: Which customers are likely to default on loans? Which are likely to be profitable, long-term customers?
Getting the right answers is important because it has a direct effect on the bottom line. Banks need to anticipate and satisfy the changing needs of the customers with a wide range of products, such as credit cards, mortgages, home equities, lines of credit, savings and checking accounts, insurance and investment products.
In addition, financial institutions require capabilities for risk management and regulatory compliance like Basel II and mandatory capital requirements. Here, banks demand best practices for decision-making in all areas of operations, including:
- Customer acquisition and retention,
- Sales and service improvement,
- Pricing and ROI analysis,
- Risk management and fraud prevention,
- Financial flow, valuation and forecasting, and
- Regulatory control and compliance.
Traditionally, financial institutions invested money and effort in predictive and descriptive models to understand key business influencers by analyzing the daily business operations data. This approach was also used to design reports and executive dashboards to understand risk and fraud, determine marketing ROI and improve business operations.
Due to slow, expensive and unreliable traditional tools, only a few business questions were addressed. With advanced analytics technology, marketing departments can now identify each market segment and create highly targeted product and service offerings for a small group of prospective customers.
Case Study: Predictive Analytics in Identifying Customer Attrition in the Insurance Industry
In the aftermath of liberalization of insurance business, life insurance has made rapid progress, surpassing even most optimistic targets. While this healthy trend is observed in acquiring new life insurance business, retention has somehow not matched this growth. Life insurance contracts are long-term in nature, and the product design anticipates the continuation of the contract for the entire length of its selected duration. The cost of acquiring new business is huge, apart from the fact that the distributors remuneration is also at its peak during the initial years. All this presupposes that unless the contracts run for their full time, it ends up as a losing proposition for the life insurers. Apart from making the business economically nonviable, attrition also results in damage to the reputation of the insurers. And, failure to keep the contracts in force ends up in a huge loss to the policyholders as well.
The customer, a leading life insurance company aiming at effectively countering policy dropouts, initiated a drive to develop a predictive model based on demographic, policy and payment variables, as well as agent-related information. When deployed, this information should automatically detect the high valued policies that are likely to lapse so that appropriate retention strategies are used.
To accomplish this, a mining structure model in SQL Server 2005 was used to deploy the model from the database to the tool environment by writing appropriate queries. This included a mining model tab where several predictive analytics steps, such as logistic regression, decision trees, neural networks and many more data mining algorithms could be tracked for the predictive decision process.
After the accuracy of the mining model was tested and showed that models were satisfactory, the model could be used to predict the likely lapses in the existing or new data sets. The results identify potential customers by indicating whether the customers’ policies are lapsed and the probability of the prediction being correct.
Predictive models analyze past performance to assess how likely a customer is to exhibit a specific behavior in the future in order to improve marketing effectiveness. Predictive analytics also encompasses models that seek out subtle data patterns to answer questions about customer performance, such as churn prediction, fraud detection and propensity to buy additional products and services. Predictive models often perform calculations during live transactions in order to guide a decision. However, the use of predictive analytics requires specialists who understand both the specific mathematical techniques and the business problem to be able to apply the appropriate techniques.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access