LOOKING INTO THE CRYSTAL BALL – PREDICTIVE ANALYTICS IN THE FINANCIAL SECTOR

By Southard Jones, Vice President of Product Strategy at Birst

Predictive analytics refers to a group of technologies that help companies infer what might happen in the future, and their own role in helping that future come about. It comes from the area of business intelligence (BI), and it is essential for financial institutions that want to move beyond the reporting and management information that most companies have today.

Whereas previously, financial services organizations might have been looking for more insight into what has happened and how that took place, the emphasis now is on what might come to pass. More importantly, this use of data can show untapped areas for growth.

More and more applications are becoming “predictive” in their approach.IT market research company IDC predicted that growth in applications incorporating advanced and predictive analytics, including machine learning, would accelerate in 2015. Apps with “predictions inside” will grow 65 percent faster than apps without this kind of functionality, according to the firm’s FutureScape for Big Data and Analytics.

Predictive analytics uses data from across a business to help organizations recognize what is taking place and their own role in influencing these situations. For financial services companies, the main target for this use of analytics is around sales opportunities. In investment management and work with high net-worth individuals, predictive analytics can reveal patterns of purchasing around financial products that can then be used to target the next suitable groups of prospects for further sales.

In practice, this means looking at past data to make recommendations for future activity. In the example of sales for investment banking services, this can involve analyzing a range of information that is both held by the company and gathered from external sources. This use of external data is important. It provides better insight to the team using data, and it also ensures that the results show more of the “big picture” and real world circumstances, rather than just reflecting the current successes that the company has had.

Use of data for predictive analytics relies on the ability to bring lots of sources of data together to build up a better picture of the existing customer base, their interests and backgrounds, and the products that they have. Based on these sources of information, it’s possible to see which customers are the most likely candidates for future financial products. These can then be targeted more specifically, rather than casting the net wider and with less chance of success.

For example, when a company is looking to sell retirement investment packages, as opposed to more general investment services, conventional wisdom might be to look at age and area of employment. However, this on its own won’t provide much help. Instead, it’s important to look at other financial products that the customer has and the socio-economic circumstances too.

This kind of analysis can provide insight into the existing customer base and where there might be greater opportunity for success around selling a product. In this example, using a combination of location data, previous products bought, age and economic data can help show where there are candidates without retirement coverage and where there is a lack of other financial products previously bought that would meet the customer’s need. These people would also live in areas where there is a stronger likelihood of being able to buy the product, making them more likely to convert into business compared to random selection.

This is a fairly simple example of how predictive analytics can help sales professionals target their activities at those who are more willing to buy. This “low hanging fruit” can help ensure that use of predictive analytics starts to get accepted within the organization. However, it’s important to look beyond this into how predictive analytics can make a longer-term impact on how the business runs its operations.

One approach is to continue modeling different products and provide that information out to the business. After all, financial institutions will have multiple products and teams that can benefit from more insight into who the best potential customers are and where to find them. However, this only prolongs that gathering of low-hanging fruit – while it is useful, it doesn’t provide long-term advantage.

A second approach is to give access to analytics across the organization for people to try out their own thoughts and ideas around where there might be correlations in the data. This can help sales, marketing or operations professionals see for themselves where their efforts are most likely to be successful. However, it should not be a piecemeal exercise where each team works with data in their own ways and their own tools.

The problem with this approach – if not governed centrally – is that each team will hold their own “version of the truth”. At best, this means that valuable ideas and opportunities will remain siloed within the team, while other employees will remain in the dark around how to use this information to best effect. Alternatively, the lack of standards in approach will mean that everyone has their own data, and the argument revolves around who is right, rather than how best to move forward.

To make the most of data – and provide the best results from predictive analytics over time – it’s important that there is a central team responsible for managing the data that is used for analysis. This ensures that best practices around information management are followed, and that all the data sources used have the right correlations in place between them. At the same time, users can be given access to data discovery and visualization tools that they can make use of for their own purposes.

This ability to bring users into the company’s data – rather than prescribing how data should be used – is an important distinction, as it will help users make their own choices around the questions they ask and the analytics they run to answer those questions. Central reporting and dashboards will remain in place as useful tools to the whole business, but the freedom to work with data has to be placed in the hands of everyone to deliver the best results.

At the heart of this is a fundamental change in how BI technology is considered – that use of data should be for everyone, rather than for the central management team. As more companies seek to become “data-driven” in their approach, it is this ability to improve decision-making for all employees – rather than just the board or the business unit leaders – that will provide the best returns. Rather than improving one decision by 100 percent, BI can be deployed to improve each and every decision made by employees, every day, by 1 percent each. The end return from BI, when used this way, is predicted to be far greater.

Comments are closed