Modern trading requires real-time data warehousing


‘In the old days’ stockbrokers used to look at stock prices at the stock exchange. They had to watch multiple screens and check their surroundings. These days the trader can manage with just one screen in front of his nose. This does not mean the information is coming from one source. He has to take it out of multiple actual sources. His boss often only has insight into his action in intermediate and end-of-day reports or sometimes through a control system that can best be described as a ‘homemade brew’.

Time for real-time

This is no longer possible in a sector where amounts and speeds are so high. It is time for the next step in data warehousing: real-time. A few stockbrokers have already made this next step. Why is this step so necessary? Here are the three main reasons.

Useful for traders

It is more useful for traders. They work with current information. They need to obtain it from multiple sources and process it in their heads into information that can be used to make decisions. Clever, but it can be done easier. By combining information in a real-time data warehouse this data can be used better. For the trader, but with the new possibilities of real-time data warehousing also possibly for a computer system.

More insight for managers

The second argument is that real-time data warehousing gives more insight into the actual situation to the managers, at every moment of the day. Traders base their actions on real-time information. Managers should therefore also keep up with the situation on a real-time basis, separately for the traders but also as a whole for the teams or departments. That is how they can, for example, identify irresponsible company risks through excessive positions and that is how fraud can be detected more easily. For this the information received by the traders and the (results of the) actions of the traders need to bundled together into one system: a real-time data warehouse. The next step is that a real-time business intelligence solution unleashes its calculations and analyses on it.

More flexibility for companies

Through the use of real-time data warehousing a trader can leave more of the thinking to the computer when it comes to combining and interpretation of all incoming data and even when it comes to converting this data into action. This allows brokerage firms to be less dependent on the individual trader. They can find people for that position with more ease and will not have to panic whenever one of their employees leaves.

Are you ready?

Real-time data warehousing has many advantages. Its branch-wide introduction is only a matter of time. Is your business ready? And how do you really start with real-time data warehousing? That is a tough question. Because right not several developers are working hard to develop or perfect applications for real-time data warehousing. Some products are already available, but these are not used on a industry-wide basis yet.

Six suggestions

Do you think you have found a reliable supplier? Keep in mind that there is a lot involved with the development and implementation. Here are a few things your organization should not be overlooking:
‘Point in time consolidation’ is necessary, which means that information always needs to be present in one of the systems if an information request is plotted. It should not be possible that ‘traveling’ data between two systems are overlooked.

  • For optimal flexibility and availability, asynchronous data exchange on the basis of ‘push’ is the most suitable. It prevents waiting for the slowest system, ‘the weakest link’ and unnecessary taxation of the trade system. Just a short amount of time loss can cost millions in this industry.
  • Uniform your definitions. Often multiple interpretations of terms like rate, exchange rate, gain, loss, and margin are possible; make sure there is no misunderstanding here. Use a reference model. Do not reinvent the wheel. Reference models also offer the possibility to answer unforeseen questions that you have never asked yourself, but other users have.
  • Take precautions to note deviations and determine trends. Deviations and trends are important signals at which time action can be undertaken.
  • Errors are unavoidable, but can be disastrous. Try to signal these directly through an iterative approach.

Guido Leenders is director of software developer Invantive that specializes in financial reports for financial institutions.

This article is also available as PDF file

Publication Banking Review October 2006