This article originally appeared in the April 2011 issue of Customer Interaction Solutions

If agents represent the contact center planet, then its core is workforce management (WFM)–forecasting, staffing and scheduling, strategic planning and performance optimization. The ability of WFM to predict and assign the right numbers of staff when they are needed and make adjustments based on deviations from plan assumptions and forecasts drives up customer-retaining service,  sales and accounts collected and avoids expensive under-and-overstaffing.

And if WFM is the core, the elements that give it its power are mathematics, statistical probability models and the underlying algorithms. The more accurate, they are the greater the likelihood that the contact centers who use them will achieve optimal service levels, balancing agent availability with costs.

ac2 Solutions (www.ac2solutions.com) is WFM to the core; the firm, founded in 2001 by Dr. Turgut Aykin and his partners, offers  what is believed to be the most advanced  WFM solution set on the market. ac2’s banner web-based Advanced Workforce Optimization (AWO) Portal, now on version 2.5, is built on a proprietary core mathematics engine that is highly advanced yet is easy to use.

The need for accuracy

Modern WFM methods had its origins more than 20-25 years ago. Initial research work was undertaken by Bell Laboratories in the late 1980s to develop a highly accurate WFM solution for its then-owner AT&T (News - Alert) to schedule the carrier’s operators. From that effort, AT&T launched NAMES10, which was one of the first commercially available WFM products in the early 1990s.

Dr. Aykin, as a professor at Rutgers University’s Graduate School of Management, was at the time researching into WFM and collaborating with the researchers at Bell Laboratories. Yet the solutions he and his colleagues saw in the marketplace were, and still are, ineffective.

The WFM forecasting solutions then generally relied on the weighted moving averages method, and some still do. The weighted moving averages method was first used in the 19th century. They also use Erlang C for staffing levels, which was invented by A.K. Erlang, a Danish mathematician, in 1917.

While fairly simple to apply, these methods are outdated and have major flaws, explains Dr. Aykin. The weighted moving average has poor forecast accuracy due to its reliance on simple weights rather than rigorous statistical models that capture trends, day of week and week of year patterns and correlation in the data accurately.

Erlang C assumes that callers will remain in queue–which may had been true in 1917 with telephone users at the time willing to wait for a long time before being connected, but definitely not today, especially in the “want it now” environment. When Erlang C is used for staffing, as Dr. Aykin points out, customer abandonments lead to significant overstaffing and inaccurate service levels, abandonment, ASA and occupancy predictions. Erlang C, or a crude simulation used by some other WFM software, is also totally inappropriate for agents with specific skillsets to enable skills-based routing. Currently there is a lot of research into this issue but there is no known Erlang-like model that can capture the complexities of skills-based routing environments properly.

The forecasting accuracy with these tools had become so poor, recounts Dr. Aykin, that the DMG 2010 WFM Benchmark study reported that the majority of users surveyed rated their WFM software’s forecast accuracy at 3 or below on a scale of 1-5, with 5 being highly satisfied. Almost a third of the participants use spreadsheets or some other software for forecasting.

“This is one of the reasons why firms do not rely on traditional WFM solutions’ forecasting functionality,” Dr. Aykin points out.

The WFM solutions also suffer from poor scheduling efficiency. The DMG 2010 WFM Benchmark study reports that majority of users surveyed also rated their WFM software’s quality of schedule optimization at 3 or below. These tools use heuristic incremental scheduling or “rule of thumb” algorithms to generate schedules. A heuristic method represents a hunch, a best guess that he says is not an effective way to achieve optimal schedules.

“Our customers have told us that they were not happy with their WFM software’s forecasting and scheduling,” says Dr. Aykin. ”The poor performance provided by existing WFM tools was the main reason for us to found ac2 Solutions. Our objective was and still is to research, develop and provide the next generation WFM and performance optimization technologies.” 

ac2’s Solutions

Dr. Aykin and his colleagues got to work on developing the next generation WFM technologies in the early 1990’s. The AWO product they came up was first released in 2004, which utilized advanced time-series forecasting algorithms including Box-Jenkins ARIMA, one of the most powerful forecasting algorithms known today, Holt-Winters with additive/multiplicative seasonality and various trend models. It also has more advanced methods such as Transfer Function/Dynamic Regression, which also analyzes call drivers to provide future looking forecasts and superior accuracy. The AWO Portal combined them with an artificial intelligence (AI)-based expert system that automatically chooses for its customers the forecasting method that optimally fits their particular environments.

“One might say that these methods are so advanced that you need a statistician to run our software, but with AWO’s the expert system, expert knowledge comes built into the software and any contact center manager with minimal training can run it from day one,” says Dr. Aykin.

AWO overcame the agent requirements forecasting flaws in Erlang C with queuing models capturing abandonments correctly for non-skills based environments. It also has discrete events simulation that accounts accurately for caller abandonments, and captures and models the complexities of skills-based routing configurations.

AWO eschewed the heuristic incremental scheduling method used by other WFM software for scheduling with ac2-pioneered integer-linear-optimization (ILO) models and algorithms. ILO captures all scheduling rules including work and off day requirements, break periods and windows, and identical or variable shift or break start requirements with hundreds and thousands of equations and inequalities. It then solves them to optimize start and break times and on-and-off days for all agents concurrently.

ac2’s AWO Portal has steadily been gaining loyal customers over the past 10 years thanks to impressive results. Their clientele have seen from six percent to as much as 13 percent drops in staffing costs with the software over the other WFM software in side-by-side trials (see Delta Air Lines cases study in the January 2011 issue).

“We’ve had Fortune 500 companies that did side-by-side studies, and they saw and understood the efficiency gains in their schedules provided by our software,” reports Dr. Aykin.

ac2 was one of the first WFM vendors to bring these applications into the cloud. This delivery method lowers IT costs while making it easier to schedule and support both in-house and the growing numbers of home-based and remote agents (see related article this issue).

“If the software architecture is old client-server, it is significantly more costly to maintain and upgrade, and more difficult to support home and remote staff, as they need to schedule their work, time off or swap scheduled work periods and communicate with their team,” says Dr. Aykin. “Web-based applications like ours only need a browser, Internet access and a password.”

Yet if ac2 has such advanced tools then why are contact centers using outdated methods, and why are competitors continuing to use them?

“The answers lie in the fact that WFM software is a very big investment, and changing them out for many companies is not a feasible option,” replies Dr. Aykin. “Fortunately with our solution, as it is web-based and it is also modularized, they do not have to rip-and-replace. They can run on our engine with the data in the legacy WFM.”

ac2’s strategy and developments

ac2’s strategy has been from the start, and continues to be, focused on its best-of-breed core technology. To help organizations improve performance, ac2 has been developing and releasing new solutions over the years. For example, almost all WFM software currently offers a performance management module. However, they are simple reporting or monitoring tools such as dashboards, scorecards, reports and real time adherence; these capabilities alone make performance management passive and not proactive.

“Performance management or accurate performance optimization involves more than reporting and monitoring,” Dr. Aykin points out. “For example, it involves optimizing skills and contact handling preferences, breaks, and overtime/VTO (volunteer time off). These functions are part of the intelligent Performance Optimization provided by ac2’s AWO Portal.”

ac2 came out with an integrated Strategic Planning module in 2009, which includes hiring, training, new agent ramp up/down, training planning, support staff planning, overtime/vto and vendor capacity planning, using the same advanced forecasting, staffing and scheduling algorithms as in the AWO Portal. This offering permits consistent and reliable results, as opposed to inconsistencies due to relying on third parties for strategic planning software and trying to plan their assumptions and modeling; these are different than that of the WFM software being used which operationalizes these plans.

The firm has steered clear of the bundling WFM-into-WFO route of its competition. Bundled solutions, says the ac2 CEO, almost universally do not offer state-of-the-art workforce management technologies.

“Bundled solutions are addressing very old concerns–from the early 1990s–in that interfacing applications was  a big problem, and because of that, companies wanted then to get all-in-one integrated solutions,” says Dr. Aykin. “Today, interfacing is not a problem. Also, the cloud-based WFM offerings are clearly offering a proof that interfacing, even over the Internet, has been streamlined to a point where it can be offered out of a box.”

ac2 has developed a back office WFM application with its partners. It can capture the unique nature of back office work and the information needed to enable efficient scheduling and forecasting.

To assist firms achieve optimal forecasting and scheduling in a cost effective manner, ac2 also offers hosted and managed WFM services. Growing in popularity since the service’s launch in 2008, with managed WFM services, ac2’s dedicated experts run the AWO Portal on ac2 hardware for clients and become a part of the client’s contact center management team to deliver needed workforce analysis and plans. 

ac2 is continuing its research work into e-learning, quality management and analytics. It is planning to release new modules later this year to provide some of these modules.

“We at ac2 Solutions are proud of the advanced technologies that our R&D team has pioneered over the years. They are providing our clients sustainable and differentiating service and cost advantages as part of our AWO Portal,” Dr. Aykin points out.