Bank Of Montreal Mines Knowledge From Data

Welcome Guest. | Log In| Register | Membership Benefits

September 20, 1999 (8:34 AM EDT)

By Mo Krochmal,

Jan Mrazek, the chief business intelligence specialist for the Bank of Montreal, is spearheading the bank's business intelligence efforts.

That process uses databases to derive deeper knowledge about customer behavior and to make strategic business decisions by combining terabytes of information. TechWeb's Mo Krochmal talked to Mrazek about the application of this technology in business.

What is the difference between business intelligence and knowledge management?

Some people call themselves knowledge management and we call ourselves business intelligence. In our particular case, knowledge management is managing something you have. Business intelligence is more proactive; you have a variety of different types of data. I can link to quantitative data, especially in the database, make it retrievable based on some relationship with quantitative data. You have to start with quantitative data and bring it together somewhere.

What are the lessons you have learned doing this?

The first lesson is to think in terms of integration. It is very important to get out of building something isolated. Integrate with systems already in place, the PCs, even the mainframes and the legacy systems. When you are designing a data mart, design it so that you can satisfy data-mining needs as well.

In designing the database, remember, performance matters. You can design the database to store the data in many different ways, usually many different wrong ways, and that leads to performance problems. A good design of a database will prevent performance problems and will allow you to grow the system and make it flexible for other unforeseen applications that come in the future.

What is the purpose of business intelligence?

The goal is not to just squeeze more dollars out of existing customers and get potential customers in the same loop and maybe get rid of unprofitable customers. I don't believe in these concepts. I believe in building customer relationships.

Who uses this system at the Bank of Montreal?

The majority of our customers are the bank's business analysts who would have to analyze this data anyway, getting it from other systems. This is now their system and they try to use it to its maximum potential, coming up with new ideas and presenting the bank with new business projects. They are getting an understanding of the overall design and they are asking for more and more. Our bank has the feeling that almost everybody could be turned into business analysts, or part-time business analysts. Anyone who feels the initiative, who has the professional interest, could start digging, learning more about the related business, the relevant information.

How do you get this knowledge?

You have different kinds of data -- behavioral data, your transactional data that will be used for behavioral analysis. You are 100 percent entitled in Canada and Europe to use this data for analysis. That is the most precious data you have if you have the robust systems, the concepts, the methodology, and the skills to use it.

Privacy laws apply to the demographic information you keep. If a customer applies for a loan, ask a customer to fill out who their employer is, where do they live, etc. The laws are concerned with this information. That is relevant, but second-tier for really advanced analytics. In our learning, we try to use behavioral data the most.

What are the privacy standards for the bank?

The Bank of Montreal is proactive. Whatever rules in terms of privacy exist, the bank has an internal privacy council, which is always much more restrictive than the laws. I'm part of the privacy council, which is 40 people from different areas. The bank is in business because it has the trust of the public, you cannot lose this trust. The primary use of business intelligence is to understand the customers. How else can you develop new products and new ways of communicating with your customers?