Usa World Bank: Scenario Analysis

USA World Bank: Scenario Analysis

March 27, 2007

USA World Bank: Scenario Analysis

USA World Bank (UWB) is a major bank with both an international and domestic presence. UWB has a large consumer and small business base and each year rolls out one new profitable product that will also increase market share. The newest product that has potential to be profitable and increase market share is a credit card that would work like a frequent flier program, where customer can earn rewards as a result of their purchases with the card. Likewise, the small business unit has also been looking at a similar program that also has a credit line of $200,000 that gives the small business owner the ability to purchase capital items. Some statistical analysis has been done on both products, but there are several issues complicating the proposal plan.

Statistical Issues

The first major barrier facing UWB is a deadline to present a single product for review by the Board of Directors. The analysis of collected data has been minimal due to the trust that Mary Monroe, Vice President of New Product Development, has for Best Market Research, the contracted research firm. Mary is convinced that the new consumer card will be the best product to present to the board. Jim Wilson, Vice President of Marketing Development, feels the new small business card would be best for Board review. In the past, the board has only allowed one product to be rolled out per year. Mary and Jim are both pushing to get their products reviewed by the Board, but with the time constraints, both teams have skipped some important analysis that would help them formulate a cohesive plan to roll out one or both products.

A second barrier facing UWB is the lack of best practices when it comes to research methods and the identification of new products. The small business unit has conducted their own research while the consumer group contracted a third party to conduct research. Neither group was aware of the analysis being performed by the other group. Likewise, UWB did not have a best practices guideline that would set research standards and give clear guidance as to how a new product should be developed.

The third barrier facing the small business and consumer groups at UWB each group's ability to understand statistical concepts as they relate to data collecting and data analysis. The senior executives at UWB are also relying on their teams to verify the validity of the data without fully understanding the data themselves. The senior management team is seeing positive results when they should be analyzing the data to see if the results truly predict a positive outcome. Essentially, the executive team seems to be making the data fit the desired outcome.

A fourth barrier is a lack of communication between departments within UWC. The consumer group is competing against the small business group to present a plan when both groups might have good ideas for new products. Mary appears to discount the benefit of the small business card because she is so focused on making sure the consumer card is the product of choice for the Board of Directors. Mary first makes her appeal to present the new consumer product to Brian while Jim is soliciting feedback from other executives. Jim and Mary seem to be positioning their prospective products politically rather than talking to each other about the return on invested capital or the ability to increase market share by launching one or both of the new products.

Additionally, When Jim presents his small business idea to Brian Allen, President of New Product Development, Brian tells him he will review the data and get right back to Jim. Seventeen days later, Jim gets a memo from Mary saying the consumer card would be the product of choice. Brian did not acknowledge the validity of Jim's data, or give the reasons he chose to move forward with the consumer card.

Bea Hansen, one of the newest members of the UWB Board of Directors has pointed out some potential problems in the statistical data presented to the Board. For instance, the consumer group had Best Marketing collect data by conducting an online survey. Online surveys can be limited to people with Internet access. Internet users are not representative of the general population and this can be a serious problem. Bea noticed that the survey was more heavily weighted with men than women, which is not consistent with the total population. Additionally, Bea points out that the error may be a sampling error which is of great concern.

Best Marketing should have used a stratified random sample to survey the banks customers. Additionally, Best Marketing should have surveyed with multiple tools such as phone, Internet, and on-site personal surveys. By do so; Best Marketing would have captured customers in various age groups, income levels, and technological levels.

Another issue with the sampling was the time frame of the sample. Sampling should be conducted within a specified timeframe to ensure the data is relevant. Consumer trends can change quickly and data samples that are collected over a long period of time can become irrelevant or invalid. In this case, Best should have checked for variability within the sample. For example, Best should have checked for variation in the sample data compared to the population from one year to the next.

Next, Bea points out that the correlation numbers are very low. The correlation between age and changing banks is three or four percent and the correlation between changing banks with the top three benefits and age is marginal at 20%. She also correctly points out that the correlation numbers may be skewed due to the large sample size. As the sample size increases, the threshold for correlation decreases. However, critical threshold can be reached when it is necessary to draw smaller samples out of the sample group to see if there is still a significant correlation in the data.

Bea goes on to point out the possible issues with the assumptions made with subjective responses. Many statisticians frown on using rating scales because the meaning of the scales is not very precise. The subjectivity can vary widely from respondent to respondent. Additionally, another board member points out that the questionnaire could have been better designed by asking specific questions that are isolated issues. Mixing issues can increase the variability of the answer and may skew the results.

Tom Araya, Marketing Associate, seems to have a good grasp on basic statistical analysis and survey techniques. Tom points out that UWB needs to stratify the random sample and ensure they are keeping track of negative and positive responses to make sure they

...