Staffing Maneuvers
By Julie Monahan
New branch staffing models are helpful, but managing them is far from automatic.
From customers' point of view, branch service is never quite right. They walk in on a weekday and see idle tellers chatting with one another to pass the time. They return on a Friday afternoon to find a packed lobby and an overwhelmed staff.
For at least a decade, managers have sought to redress such imbalances by using automated workforce management tools, essentially software-driven statistical models, that try to balance the needs of customers with internal efficiency requirements. Now the balance is tilting toward the customer, with staffing models used as much to improve service as efficiency.
"Staffing optimization" goes beyond basic scheduling to match service levels with forecasted customer demand under multiple scenarios. When handled correctly, the process can help ensure that branches always have the right number of representatives and the right array of expertise on hand, in situations ranging from the Friday afternoon rush to the off days.
"It's not just about improving efficiency, but rather it's about customizing capacity," says Jim Eckenrode, group director of consumer banking research at TowerGroup Inc., Needham, Mass. "Every branch is different and staffing must be tailored to particular markets."
The increasing sophistication of the exercise has been facilitated by ever more robust software. Whereas early workforce management tools simply helped streamline a manual process, staffing models now use historical transaction data to identify trends in employee productivity and customer behavior. The data is compared with proprietary algorithms and internal service benchmarks to produce transaction forecasts and specify the staff hours needed to meet them.
Proponents say staffing optimization improves branch productivity by up to 15%, reduces customer wait times and liberates managers from scheduling duties. To the extent that these systems permit staff to request work shifts that match their needs, they have the potential to improve morale and also cut employee turnover.
It sounds so, well, automatic. But accuracy is a key sticking point: If staffing projections don't align with front-line realities, the exercise is defeated. So strict data monitoring is required. Despite their analytical power, these systems still rely on certain critical assumptions, such as the amount of time estimated for a rep to help a customer reconcile an account or handle contingencies. Consultants must be brought in to validate the assumptions or branch employees themselves must track the activity.
There are political issues as well. Branch managers may fear losing control under the new system, while employees worry about losing their jobs. Senior managers must reassure both that staffing optimization helps them work smarter.
Employee Concerns
Staffing optimization, at its most basic level, is just another software exercise. Its success, however, depends on a host of other considerations that reach into the institution's very culture. Employees must learn to plan their work schedules around the activity patterns of customers, as opposed to their individual preferences, says H. Randy Lee, a senior vice president and regional finance manager at Wells Fargo & Co., San Francisco.
Managers also need to adjust their thinking. "Branches will tend to staff to their peak requirements — hiring enough staff to handle the line going out the door on Friday and then generally overstaffing the rest of the week," says Darryl Demos, chief executive officer at Demos Consulting in Norwell, Mass. "These tools give managers the courage to try new strategies."
Both branch managers and employees may resist this change — managers because they fear a loss of control and employees because they fear layoffs. Some institutions have responded by centralizing staffing decisions, essentially bypassing the branch managers. But optimization advocates insist that banks will elicit better compliance with electronic schedules and forecasts by preserving at least some level of branch manager involvement and by providing adequate training and orientation to staff.
Consulting with managers during the development phase helps instill more confidence, especially when there's leeway for managers to point out information gaps and discrepancies between automated forecasts and the day-to-day realities in the branch. "We reviewed all of the statistical assumptions with the managers in our financial centers," says Amy Shreve, a vice president and manager of financial center services for First Tennessee National Corp., Memphis, Tenn. "It was a team effort."
While branch managers might fear that the use of statistical algorithms relegates them to the background, the opposite may be true, since scheduling programs still rely heavily on each manager's observations. While software can generate a month's worth of forecasted schedules, managers still need to decide how to fill those hours based on each employee's skills and status.
And while automated tools rule out certain human errors, such as scheduling two tellers from the same branch for the same week of vacation, no system can eliminate the unexpected absence or sudden influx of customers. "These systems are dealing with averages," TowerGroup's Eckenrode says. "The exceptions are the ones that are difficult to manage."
Employee fears may likewise be overblown. For one thing, the systems are designed to incorporate employee choices by keeping track of work-hour preferences, class schedules for employees in school and other scheduling needs. "Once employees know you're interested in where and when they want to work, it sends a message that you care about their preferences," Demos says.
And while it's true that staffing models can identify situations where a branch is over-staffed, the programs also pinpoint areas where more staffing is required to improve customer service. "Any bank that seeks to do this has to stay open-minded to the possibility that certain staff requirements might actually increase," says one banker, who did not wish to be identified. "We actually expect a slight increase in staffing at our bank because we did not have the correct standards in the past."
Forecasting Game
In practice, banking companies have developed a variety of criteria for evaluating automated workforce management tools.
Wells Fargo asked each prospective software supplier to make several months' worth of staffing forecasts, based on two years of data for eight branches. The results were then compared with the actual staffing needs surfacing during the forecast period. The vendor with the highest rate of accuracy in forecasting won the account.
The company formerly set expectations for the number of transactions per branch and applied that benchmark network-wide. But that benchmark often did not match the volume of business at smaller branches, causing them to over-staff in certain situations. After incorporating its chosen staffing model, Wells Fargo cut the work hours spent on scheduling by nearly 90%, according to Lee, who says customer service also improved. "We can predict when a customer is going to be there and how long the transaction will take," he says. "We don't use rules of thumb anymore."
First Tennessee focused on the level of detail in the forecasting system as well as reliability. Shreve says staffing models are distinguished by their ability to assign different values to different types of transactions. Some packages "consider that transactions have the same resource requirements — whether it's cashing a check or taking a commercial deposit — though in reality they don't take the same amount of time."
After replacing the manual scheduling system it formerly used to manage 1,000 tellers, First Tennessee trimmed labor costs by reducing overstaffing during the mid-week and adding more part-time staff on busy Fridays, Shreve says. "We found that our financial centers were scheduling as if every day was their peak."
UnionBanCal Corp., meanwhile, looked for yet another feature in its staffing optimization model — an ability to accommodate the service needs of different ethnic groups. Customers for whom English is a second language tend to take their time in branch transactions, partly because of language barriers and partly in the interest of handling transactions carefully and thoroughly.
"Our transaction-time standards are based on expectations for that particular marketplace," says Ronald Kendrick, an executive vice president based in San Diego. Following pilot tests in 29 sites, Union Bank is rolling out its system to 200 branches.
Given all the belt-tightening and layoffs that occurred in banking during the past decade, not many institutions are likely to realize significant labor savings with new staffing models. Instead, bankers and experts agree that improved customer service is the real benefit of staffing optimization. "Better peak-time staffing is inextricable with better service," says consultant Gordon Goetzmann, a Seattle-based managing vice president at First Manhattan Consulting Group.
Ms. Monahan is a freelance writer based in Seattle.