Ask the Expert

Response rates vary tremendously from one industry segment to the next, and even within an industry segment. A 50% response rate is very good, but it can range anywhere from as low as 15% to as high as 70%. A very low percentage can have various causes:

• There are too many questions in the survey
• The customer has been surveyed recently
• The respondents do not use or recognize your services

Sometimes a low return rate indicates a low level of satisfaction. In a sense, your response rate will depend on the relationship you have with your users. With a global customer base, you may see culturally-based response rates. For example, your French customers may respond infrequently, while your German customers may respond enthusiastically. I recommend that you discuss with your colleagues in other countries what their attitudes are toward completing surveys.

I always recommend that when dealing with the technical aspects of security, you work with your IT security people and the vendor you bought your tools from. There may be aspects to products that they can help you with.

As for supporting VIPs, you are only limited by your imagination, time, and budget. The most common techniques are:

• An e-mail code that VIPs would place in the subject line of an e-mail message to support, indicating that service is required ASAP (this technique is most effective when you monitor your support Inbox closely),
• A dedicated “red phone” line for VIPs (almost always very popular), or
• A hidden option on the ACD that allows for immediate access to the support center.

You should also have a support plan for each executive that includes a list of the equipment they use in their office and when they travel.

There is nothing definitive that says that your approach is the correct one; however, commonsense, concern for customers, and the spirit of Section 404 in Sarbanes-Oxley and COBIT (Control Objectives for Information and related Technology) all suggest that any activity related to security and compliance be restricted to as few organizations and people as possible. Your approach appears to be working, and is good IT governance as well, so I suggest that it not be changed. If a change is forced upon you from the “other side,” then they must be made to sign an ironclad operational level agreement (OLA) that makes them adhere to the standards of efficiency, effectiveness, security, and service that the organization is now providing. Finally, if any mistakes are made or there are any problems with the new applications or logins, users will call the service desk anyway and the service desk will have to contact the business owner. That’s very inconvenient for the customer and ignores their needs. It doesn’t make much sense to me; the business owners should learn to play nice with you! Tell them I said so!

Beyond the applicable federal (Fair Labor Standards Act) and state labor laws, you can do as little or as much as you want for the holidays. I’ve been in organizations that were very generous and one’s that were very cheap; it depends on how much your organization values its workforce.

One approach would be to check with your nearest large hospital and ask their HR department how they compensate employees for the holiday work and follow their lead. Keep in mind that there will be a difference in union and non-union compensation rates. Personally, I’d go with the union compensation rate; it is a nice gesture and sends a good message about how much you value your people.

Also, try to rotate the people working during the holidays, if you can so no one person works every holiday, unless they want to. And ask for volunteers first; it is always surprising how many people like to work during the holidays. This period tends to be fairly quiet and more relaxed then regular days. In addition, consider supplying food and beverages for your holiday workers; again, not required by law, but a nice gesture that sends a positive message.

You’ll be glad to know that your situation is common. First call resolution is what I call a “legacy metric” from the call center industry. It has been around since the very first call centers were created in the 1970s. It was defined then as the percentage of calls that solved a customer’s problem or answered a customer’s question on the first call. For more service-oriented calls (e.g., credit card bill inquiries, catalog orders, burnt out street lights, etc.), it was an excellent KPI to watch. Fast-forward to 2010 and we find that technical support operations are being held to forty-year-old standards. No one wonder managers are losing their hair!

FCR is just one key performance indicator (KPI) among dozens, and it has been endowed with far more importance than it deserves. If it is too high, it suggests that the calls the support operation are receiving are either very easy or what is being measured is very limited; cherry picking is never a good idea. FCR is merely one indication of a support operation’s performance; it is like blood pressure, pulse, and respiration—an indication of health. Just as you would not want your doctor to “fudge” these indicators, I suggest that the first thing you do is to stop parsing what you measure. If you don’t count everything, FCR is useless. If you are receiving calls about parts of the IT operation that you don’t control, then record the calls and track how they will be resolved, when they are resolved, and how often you receive these calls. Finally, don’t take FCR so seriously. Unless you have control over and can influence every aspect of the IT infrastructure, it’s just not that important. Look at it in the context of your entire operation and explain to your employees that FCR is just one KPI among many.

The ratios you are describing fall under the description of span of control and there is no “one size fits all.” The manager-to-staff ratio depends on a number of factors, such as:

• The manager’s level of experience
• The manager’s organizational skills
• The manager’s leadership style
• The support analysts’ maturity
• The customers’ needs
• The manager’s job responsibilities (i.e., very busy managers can manage fewer people)

The “Art of Software Support,” by Françoise Tourniaire and Richard Farrell, suggests that 9–12 is the average number of managers needed, but like all such rules, it depends on too many factors to generalize.

I recommend these three books:

• Call Center Management on Fast Forward: Succeeding in Today's Dynamic Customer Contact Environment, Second Edition (ICMI Press, 2006), by Brad Cleveland: Covers forecasting and scheduling for call centers
• Managing and Motivating Contact Center Employees: Tools and Techniques for Inspiring Outstanding Performance from Your Frontline Staff (McGraw-Hill, 2002), by Malcolm Carlaw, Peggy Carlaw, Vasudha Deming, and Kurt Friedmann: A long book, but a good one.
• Wake Up Your Call Center: Humanizing Your Interaction Hub, Third Edition (Ichor, 2000), by Rosanne D’Ausilio, Ph.D.: A quick-and-easy guide to call center customer service.

As to KPIs, your KPIs will depend on your business environment and what is important to your business and your customers. At the very least, you will probably want to measure three categories of metrics:

• Response KPIs
• Resolution KPIs
• Customer service KPIs

The best way to determine the number of analysts you need is to base staffing requirements on business needs, service level objectives, and customer expectations. It’s not as philosophical as it sounds. Here’s what you should do:

• Determine what your service level objective will be. For example, “Answer 80% of calls in twenty minutes or less, with a talk time of seven minutes and a wrap-up time of two minutes.”
• Use an Erlang C calculator. Insert your numbers into different boxes, hit a button, and it will give you an estimate of the number of staff you need. The Call Center School (www.thecallcenterschool.com) has a very nice Erlang C calculator called QuikStaff that you can download for free, as well as several very well-written articles on staffing and scheduling.

Since call volumes can vary from half-hour to half-hour and hour to hour, you will have to build a staffing model for every half-hour, every day of the week. I know, it sounds awful, but the ratios you asked about are about fifteen years old now and were based on a Gartner report that has not been kept current.

The approach I always recommend is to use an Erlang C calculator and a recognized staffing development methodology. You can find an excellent, free Erlang C calculator at www.thecallcenterschool.com in the Tools section; The Call Center School also has several excellent papers on staffing and scheduling.

As far as shifts go, it depends on what kind of service you want to deliver and what your service level objectives are. Generally, however, eight hours is better than twelve because people get tired at around ten hours and you’ll always need two people for each shift to allow for meals and breaks.

Your problem is one of service level management; it just doesn’t seem like it. You are correct in assuming that when your customers do not receive a “timely” telephone call or e-mail, it has a negative impact on how your operation is perceived. It is also common to measure various aspects of how this relationship is managed.

What you are looking at is a series stages in the transaction between your operation and your customer:

• Work order arrives and is entered into management tracking system—date- and time-stamped
• Work order is assigned to a technician—date- and time-stamped
• Technician makes first contact with the customer—date- and time-stamped

Each of these stages should be governed by a service level objective and the management tracking system should be updated throughout the process. One question is, who will manage this process? Ideally, you should design a process that will run automatically, but that rarely happens so you may need to appoint a service level manager. I highly recommend the ITIL material on service level management for more detailed guidance.