“Thank you for calling the service desk. What’s your emergency?” OK, so maybe the traditional 911 operator greeting could be construed as sarcastic if spoken by an agent responding to an Outlook plugin issue, but it does put the level of urgency into perspective when comparing lightning fast response times for relatively minor inconveniences with those of a four-alarm fire. At the service desk, faster isn’t always better. For this reason, a CIO or manager will often consider the financial consequences before setting an overly aggressive Average Speed of Answer (ASA) target, particularly the less efficient use of resources to deliver on that metric. Certainly, a five second ASA is achievable and certainly any interruption in fully functioning IT assets can feel like eons to the end user awaiting live agent support, but staffing for that target on a 24 x 7 x 365 basis can drive up costs unnecessarily when compared to more realistic, industry standard SLAs. While monthly reporting and analytics may vary from one organization to another, an overwhelming majority of IT support issues, barring ones that pertain to urgent patient care for medical practitioners, are not life or death. On the other end of the spectrum, interruptions in any employee’s course of work also come with a real cost that cannot be taken lightly. That’s why most organizations compare the costs and strike a balance before setting an ASA too high or too low.
If the number of agents is aligned to accommodate a 5 or 10 second ASA during anticipated “worst case scenario” peaks in inbound contacts, eventual dips lead to waste due to underutilization. Superfluous costs of underutilization are further compounded in a dedicated staffing model in which agents are on the company dime whether they’re busy or not. Only a flexible service desk staffing model can assuage the operational tug of war between staffing levels and utilization with regard to meeting ASA targets over periods of extreme volume fluctuations. Unless agents are willing to clock in and out intermittently throughout the day or work staggered shifts, a shared staffing model is a reasonable alternative for organizations willing to engage a service desk outsourcing solution. A shared staffing model incorporates primary analysts dedicated to one account as well as a pool of secondary analysts who are trained to handle additional contacts during unexpected peaks in volume due to outages, rollouts, growth etc. In this model, call volume spikes overflow via the ACD from a pool of primary agents to the remaining shared resources so SLAs such as ASA are generally not adversely impacted during periods of excessive call or contact volumes. By contrast, an internal solution is effectively a dedicated staffing model in which a finite number of agents can miss ASA targets during peak periods or fall prey to underutilization whenever demand subsides.
Regardless of whether or not the service desk is internally staffed or outsourced, the team of agents dedicated or shared, IT managers still have to determine appropriate staffing levels required to meet the target ASA on a monthly basis. So striking a balance between a manageable ASA and the agent count and costs necessary to achieve it are generally the first order of business.
Setting the number of available agents based on ASA is a straightforward calculation using the Erlang C formula which addresses random rather than sequential workloads. If a client generates 50 contacts an hour and Average Handle Time is 6.5 minutes, the number of agents required to maintain a 5 second ASA is 10. For a 60 second ASA, leaving the number of contacts and AHT as is, that number of agents drops to 8. Considering an agent may be compensated $25/hour for salary and benefits, the annual costs for those two additional agents in this example can be just north of six figures…and three times as much if coverage is 24 x 7 x 365 (i.e. beyond an 8-hour shift per agent). With such skyrocketing annual costs standing in the way of a 55 second ASA differential, organizations that ask “can we afford to wait” may no longer find this a rhetorical question. Thankfully, running in the ASA race is more of a marathon than a sprint with the qualifying part of the acronym referring to the “average” which takes into account the busiest Monday morning and the most silent of holiday nights, up and down the org chart, and for issues ranging from server outages to converting a Word file to pdf. Whatever combination of ASA and staffing levels gets the service desk to the finish line on time, without exhausting personnel and budgetary resources, will be a race well won.