The First Applications of Global Analytics
David J. Fogarty
The question you may be asking yourself is what is Global Analytics? You may already know about the power of analytics and data-driven decisions. But what the process of doing this globally and how has this led to the management of remote teams? My first experience with global analytics was in the early 1990s when I worked as a risk analyst at GE Capital.
At GE, Global Analytics was all about the management of remote centralized teams that helped a global business to grow its international startups and acquisitions via providing analytic support to these subsidiaries. There was a gap in skills across the markets which could be arbitraged via global analytics and this created huge value for global firms such as GE Capital.
This key business driver would have not been possible without the creation of remote working teams and the management processes to lead them.
One example of this for GE was the acquisition of one of the largest consumer loan companies in Japan that issued loans with manual underwriting and no analytics tools. It long been proven originally by Bill Fair and William Isaac who created the company Fair Isaac that analytics in the form of credit scores could improve the underwriting of consumer loans over that of manual underwriters. The logic behind this is that there are many variables which go into a person’s willingness to pay back a debt and that a human being underwriting a loan cannot possible calculate all of the variables and their relative weights in their heads. Therefore, a mathematical interpretation of all of those variables in the form of a score has been proven to be able to outperform human judgments. This have been proven in multiple peer reviewed studies including Blochlinger and Leippold (2006) and Fogarty (2006, 2012).
Given this the due diligence team noted that the credit limits on these consumer loans were very low partially due to the conservative underwriting. So, the premise was that if GE Capital could acquire the company and use analytics to create credit scores it would be able to better assess the risk of individual customers and raise the credit lines. Once credit lines were raised on the less risky customer then spending would go up and you would create incremental growth via charging interest and fees on the increased balances. Well the business was then acquired and my centralized global analytics team in the UK then started transferring massive amounts of data from the mainframe computers in Japan to Leeds, UK for analyzing. We also have a representative from the Japanese business come over to Leeds to help with the translation of both the data elements and also to communicate back to the business the results of the analysis. The analysts in Leeds then followed the process of analyzing the data and created credit scores. These credit scores were then transferred back to Japan and implemented on their platform in order to score all current customers and provide them with a risk probability assessment or a score. With this, the lower risk scoring customers were then automatically given increased credit lines on their accounts. When you give automatic credit line increase to revolving accounts you typically see a boost in spending as not customers have more credit to use to make purchases. In one of the largest consumer loan companies in Japan this resulting in billions of dollars in receivables and several hundred million in additional income resulting from GE Capital’s acquisition and the resulting global analytics efforts. This work partially fulfilled the goal of GE Capital of being able to acquire new businesses across the globe and then grow them using GE Best Practices. This is also a great example of where globally located and accessible teams can become part or an acquisition strategy where the acquiring company finds assets to buy which do not have advanced analytics and then apply best practices to analyze their data and grow the business. Overall, this is an excellent example of early global analytics.
Blochlinger, A., Leippold, M. (2006). Economic Benefit of Powerful Credit Scoring. Journal of Banking and Finance, Vol. 30, no. 3. Pp -851-873.
Fogarty, D., J. (2006). Multiple Imputation as a Missing Data Approach to Reject Inference on Consumer Credit Scoring, InterStat Journal, Sept. 2006.
Fogarty, D., J. (2012). Using Genetic Algorithms for Credit Scoring System Maintenance Functions, International Journal of Artificial Intelligence and Applications, vol. 3, no. 6, pp. 52-60.