Traditionally there was little call for innovation in the banking community given that financial services, while local (unlike products) were viewed as commodities. However, to combat the competition in the late 1990s, “organic growth”, the growth that a company can achieve by increasing output and enhancing sales, was the way to go, as it would reduce earnings volatility that resulted from acquiring new banks and increase profitability. Thus the I & D team was launched for product and service deployment. To risk the possibility of large scale failure, the experimentation of new technology and services was initially confined to a set of 20 (later 25) branches in Atlanta. In order to protect these branches from market fluctuations that could lead to product termination, it was suggested that these branches function as separate R&D branches, free from the day-to-day responsibilities of running a bank. However, this was contested by the fear that the results from the I&D market would not be duplicable elsewhere.
Atlanta was chosen as the site for the I & D market as it represented a stable market. The banks had the most advanced communications network and Atlanta was close to HQ in Charlotte. As a part of the I&D market, there were 3 types of branches: express centers for quick, routine transactions, financial centers with more complex technologies and access to a wider variety of services, and traditional centers, which were simply traditional branches with enhanced processes and technology. Prior to introducing these experiments, the team typically rehearsed how the activity should occur in a prototype center in Charlotte. For instance, the Walt Disney Company designed and taught them the Bank of America Spirit program which laid emphasis on increased customer relations.
There were several problems faced by the team: how to prioritize ideas, gauge success, run several experiments at once, prevent the novelty factor from affecting the outcome and defend the team from budget cuts. Experiment cycle time was 3 months to ensure rapid feedback. Due to the 18 month concept to national rollout lag as opposed to the 3 month testing period, productivity was used as an indicator of success instead of financial performance. To prove themselves and meet the bottom-line, ideas with a higher probability of success were given a higher priority. In order to isolate the effect of an experiment on branch performance, the team repeated trials, averaged the results and paired experimental branches with a branch operating under normal conditions to minimize the effect of noise factors. In order to increase the number of experiments running at once almost 15 experiments were run per branch and specialists were brought in to analyze the effect of each variable (experiment). The TZM experiment which was intended to reduce perceived wait times, in fact, decreased the number of people who