Networks in Action

Walt Coleman, the vice president of operations, who in 1989 was brought in to head the project, spoke candidly to CIO Magazine in mid-1992 about what he as a non-technical executive had learned about handling a technical operation. The goal was to respond to customer surveys that indicated that a key need for Whirlpool was to be more accessible and for its agents to be more responsive. The first step would be to implement a one-call service system-a hotline to diagnose customer problems on the spot and either provide immediate assistance over the phone or dispatch a service technician.

Whirlpool makes a million service calls and handles three million phone calls each year. Whirlpool consolidated 26 district offices into two consumer-assistance centers. The scale and complexity of the technology needed to bring together one-call telecommunications, accurate diagnosis, and reliable and responsive answers and actions were substantial:

1. The telecommunications component involves a 1-800 call into an AT&T T1 network. An area-code router sends it to either the Michigan or Tennessee center. Redundant T1 networks provided by AT&T and MCI link the two.
   
   
2. Each site has an AT&T Definity 2 PBX. The lines are ISDN Primary Rate Interface. The Definity uses a look ahead feature to route calls to the first available agent. Software in the Definity PBX passes the callers number to an IBM 3090 mainframe host, which matches it with a customer profile.
   
   
3. The data is passed to an agent, who uses an IBM PS/2 personal computer. The PS/2s are linked in work groups of 30 by IBM 16-Mbps token-ring LANs connected to a 16Mbps token-ring LAN backbone. The PS/2s also use IBM's Extended Edition software, which allows both workstation-to-workstation and workstation-to-host communication .
   
   
4. Each LAN supports an image server that houses four CO-ROM drives with around 150,000 pages of scanned product and service data covering 20 years of products. The total paper documents would stretch for 26 miles.
   
   
5. An expert system on the LAN provides a series of questions for the agent to ask the customer over the phone. It steps through a dialogue that ends either by recommending action the customer can take or scheduling a service call.

Whirlpool estimates that the expert system may reduce service calls by 10 percent; at $40 a call, the elimination of 100,000 calls represents substantial savings

Walt Coleman had no idea of just how complex the project was. The firm's Information Systems (IS) group was poorly positioned to handle it; for years it had been, according to Coleman, a "background function, involved in payroll, batch reporting, and so forth. It was to be seen and not heard". Rich Koeller, who took over as head of IS on the very same day as Coleman took over the customer service center project and who was to become a key ally, says of the old IS department, "There was no way we could do a massive project on time with our resources". Koeller is widely respected in IS circles for his character, candor, and capability; his job was to stir up IS and bring it into a new service focused, business-driven environment.

Coleman decided to hire IBM to evaluate the project and outline the technical requirements and costs. Based on IBM's report, he decided to hire the company to design and build the system. The project team consisted of 75 IBM staff and 15 Whirlpool employees. He assigned one of his own people, Bruce Dacre, as project manager. Dacre understood the business well and had some knowledge of information technology. Because IS was being reorganized, he decided not to look for a "champion" in IT to provide technical advice and guidance, but instead to move ahead as quickly as possible using the available resources and contacts.

He now views this as a major mistake on his part. He relied on IBM for technical leadership, and because the systems integrator made the technology sound relatively simple, he was lulled into believing that the emerging technical problems would solve themselves. He did not ask enough questions or provide strong direction. In this very common situation, the project team wandered off course. It did its best, which in this context meant providing the best technical design. Coleman says, "They were going for technical elegance. They were doing what they thought was right, and we were thinking they knew what we wanted. But they got off on tangents, and we did not provide the discipline to stay on course".

Without clear guidance, there is no basis for design trade-offs. For example, is 30-second response time acceptable? Ten seconds? Five seconds? How much will it cost to go from 30 to10 or ten to five? Is it worth the cost? Koeller points out that it requires an experienced implementer to understand that the cost of moving from 30 seconds to 10 is relatively low; it can generally be handled by selecting and fine-tuning hardware. But going from ten seconds to five is much more difficult and is more likely to involve fine-tuning software, this can tie up an unjustifiable amount of effort and may not be worth it.

IBM's team had to work out what technology was enough and which technical features were essential. Dacre did his best to control the project but lacked technical knowledge. Coleman accepts that Dacre was in over his head and Dacre did his best to shield Coleman from having to address technical issues so that Coleman could concentrate on the business side of the project. When a key delivery date was missed, Coleman asked for help from Koeller in IS. Koeller's judgment was that the original decision to use an IBM AS/400 mid-sized computer that was not integrated into Whirlpool's corporate platform was a mistake; the system needed a much more powerful mainframe machine. Koeller assigned one of his own experienced project leaders, Jan Massey, who worked in close contact with Dacre. Dacre now concentrated on the business side, and the IBM technical team reported to Massey. Coleman comments that "she should have been on board from day one. We need IT leadership that can understand the business ease. Conversely, we need business people who can venture into technology to get some ideas. Both need to cross the line to communicate better".

Naturally, senior management was not pleased by the unexpected news that a project that had been reported as being on schedule was now in major trouble and would cost far more than budgeted. Getting back on track was difficult and politically sensitive. "The wheels had come off and we had to get our confidence back". Massey was the key person here; combining her strengths with those of Dacre were vital to making things work.

The new team completed the project in June1992, meeting the revised schedule. Coleman summarizes the lessons he learned:

1. "I'm thinking more like [IS] and they're thinking more like me. I have a lot more knowledge and understanding of technology than I ever thought I wanted."
   
   
2. Awareness of the technical issues helps assess the business case and the trade-offs between the costs and benefits of technical enhancements.
   
   
3. Dialogue leads to deeper levels of collaboration; instead of just supporting Coleman's goals, IS helps the business group establish goals.
   
   
4. Sensitivity and collaboration cannot be developed in a vacuum: "Forget about running seminars on mutual understanding. Instead, start with a clear mission and build a team."
   
   
5. Familiarity breeds familiarity; sharing successes and failures helped the. multi-unit team coalesce.

Questions for Discussion

1. What was the business need that led Whirlpool to develop a state-of-the-art customer service center, and where would this need fit in the telecommunications business opportunity checklist?
   
   
2. Walt Coleman, VP of operations at Whirlpool, brought in one of his own people; Bruce Dacre, to act as project manager on the service center project. Was this a good move? Explain.
   
   
3. Using information from the minicase, discuss the importance of clear guidance from the users as the basis for design trade-offs. Also, comment on whether there was sufficient user guidance to develop an appropriate design for the Whirlpool service center.
   
   
4. Discuss the role of dialogue among user, designer, and implementer in the case by addressing the following:
   1. its importance in meeting the goals of the project
      
      
   2. how well the project was actually carried out
      
      
   3. ways in which the project could have been improved.
      
      
5. What lessons were learned by management about both the role of teams and the mix of skills needed on the team for this type of project)