Calculating Service Desk Staff Numbers.jpg

How to determine the right number of Service Desk staff

Numerous factors impact Service Desk staffing numbers, making industry benchmarking and comparisons very difficult. Even when benchmark data can be found, the scope of services, and even the definition of what constitutes 1 Service Desk FTE, vary considerably.

Before you read on, I’d like to warn you that there is no magic formula for determining the optimum number of staff for your Service Desk. However, I will share a few methods that are available and then I’ll provide a comprehensive list of things that you should consider when you need to justify your Service Desk headcount, justify an increase, or are looking at ways to reduce it.

First, let’s look at some of the options that are available.

Erlang C

Erlang C is a formula used by call centres to help them determine their staffing numbers. It can provide a starting point, but is designed for single task call centres (e.g. inbound customer service agents whose primary task is to handle those inbound calls).

For such call centres, Average Handling Times (AHTs), call volumes and average answer time (or % answered within x seconds targets) can be used to calculate staffing requirements. Leave, extended hours etc. can then be factored in to produce a headcount number.

For Service Desks that are based on the ITIL concept of the Service Desk function, the breadth and depth of tasks is usually far greater than simply handling inbound calls within an average handling time.

I therefore don’t recommend the use of the Erlang C formula when determining Service Desk staffing levels.

Gartner (and other research)

Doing a Google search, I was quickly able to come up with some Service Desk stats. Here’s two:

According to this thread (http://www.techrepublic.com/forum/discussions/102-210878), Gartner provided the following benchmark information in 2007:

  • 414 calls per analyst per month
  • 1.15 calls per each user into a Call Centre
  • 63.2% FCR (First Call Resolution) Rate
  • 7.07% Abandonment Rate
  • $23.71 Cost Per Call

However, another study I found (http://itbenchmark.wordpress.com/2009/07/30/help-desk-staffing-ratios-by-size-and-industry/) came up with a different benchmark by taking a different approach. This study reported:

  • “The median staffing ratio is 1.3% (in other words, 13 help desk support personnel are supporting 1,000 company employees, or one help desk headcount for every 76.9 company employees)”

The situation is again exacerbated by the scope of the support functions being studied (e.g. Call Centre vs Help Desk vs Service Desk) as well as the definitions used within the study (e.g. The definition of ‘First Call Resolution’ can vary quite widely).

There’s so much variance in the measures being used for benchmarking that its almost impossible to make a like for like comparison between the various benchmark information available. How do we know which information we can rely on?

Why is it so hard to benchmark Service Desk staffing levels?

It is so hard to benchmark Service Desk staffing levels because of all the factors that make every organisation unique.

Here’s a list of the factors that spring to mind. At least you can explain to your boss why you’re finding this task so tricky!:

  • Call volumes and their breakdown by channel (call, web, email).
  • Staffing mix (full-time vs part-time, permanent vs contract vs casual).
  • Number of public holidays, annual leave allowances, typical number of sick days.
  • Hours of operation of the Service Desk, e.g. 24×7 vs 9 to 5.
  • Required Service Levels, e.g. Time to answer SLA (e.g. 90% in 60 seconds), calls abandoned post-IVR, 1st Level Resolution targets.
  • Scope of services supported at 1st Level.
  • Degree of task automation (including call logging, request fulfilment, information gathering via web forms, automatic deployment of software etc.).
  • Additional tasks and duties, e.g. reporting, admin tasks, servicing remote outposts (e.g. library counters, walk-up desks, genius bars), lease refreshes.
  • Calibre of Service Desk staff available to you. Are you able to recruit the best Service Desk staff from the market? Pay, conditions, reputation, career prospects, willingness and freedom to hold out for the right person all have a bearing on the ability to recruit the best staff. Given that top Service Desk performers outperform average to poor performers at least 2 to 1 (in my experience), recruiting the best from the market can make a big difference in how many staff you need, or how much Level 2 work the Service Desk can take on.
  • Complexity of the environment being supported. The more diverse the environment the Service Desk supports, the more difficult it is for staff to build knowledge and provide rapid resolution to a reasonable scope of incidents. It also takes longer to induct staff in complex environments, which has a greater impact in the Service Desk than anywhere else in IT due to the higher attrition rate.
  • Volatility of the IT environment. Is there a major change underway, e.g. new SOE deployment, ERP upgrade, or roll-out of new desktop software.? If so, call volumes may taper off over the longer term as the project winds down and customers adjust to the change. Is there an upcoming major change planned? If so, this might require additional headcount on the Service Desk to field the likely increase in call volumes.
  • Factors related to your customer base, including: level of dependence on IT (some companies might have a number of staff who do not have their own PC, e.g. on manufacturing sites); criticality of IT services (if all services are critical, more rapid response to calls and incidents will be required, requiring higher staffing levels, all other factors being even); typical IT literacy of your customer base; use of Super Users or Subject Matter Experts within the organisation.

So, what can you do?

There many reasons for considering Service Desk headcounts in the first place. Are you aiming to reduce headcount? Free up Service Desk staff to do more interesting tasks? Extend the hours of operation without hiring? Defend your existing headcount? Justify additional staff?

Regardless of your goals, here are some of the things you can do to ensure you have the optimum staffing levels for your Service Desk:

  • Instead of looking for industry standards, use your current staffing levels,  typical work volumes and service level performance as a baseline to work from, and compare back to, as you look for opportunities to improve.
  • Look at opportunities to reduce incoming incident volumes  – Problem Management is overlooked in many organisations and yet it benefits the customer and saves IT costs..
  • Look at opportunities to automate part or all aspects of high volume Service Request types, e.g. password resets or desktop software installs.
  • Consider other opportunities to reduce inbound calls, such as: effective usage of IVR messages and online noticeboards to head off calls about known outages; creation and promotion of FAQs and Knowledge Management articles.
  • Consider how task workloads and staffing levels are spread and structured to allow for peaks and troughs in customer demand. One example is to ensure staff are working on emails between calls.
  • Utilise technology to give Service Desk Analysts different priorities in the phone system to help simplify the management of peaks and troughs (e.g. your primary queue manager and primary email manager only get calls when everyone else has been busy for 30 seconds and calls are waiting).
  • Look at other tasks the Service Desk currently fulfil. Should they be doing these tasks?
  • Look at the impact of walk-ups / interruptions to the Service Desk. Is the organisation prepared to wear the cost? Would Genius Bars or scheduled office walk-arounds be better?
  • Consider the cost/benefit of attracting higher skilled staff into Service Desk positions.
  • Consider engagement levels of Service Desk staff. Higher levels of staff engagement have been shown to postively impact producvity, increase customer satisfaction, reduce sick leave and increase retention.
  • If you need to reduce headcount, can response and resolution targets (time-based service level agreements) be made more lenient or even dropped altogether? An over focus on traditional time-based SLAs can result in green SLA traffic lights while customers are still unhappy. Conversely, SLAs could be showing amber/red while customers are happy. The bottom line is that your SLA traffic lights might not bear much relation to how your customers perceive the service they’re getting. So there may be more room to adjust headcount than you think.
  • Finally, are you measuring customer (user) satisfaction with your support levels on a ticket-by-ticket basis? If you are measuring satisfaction levels (and using customer feedback for staff coaching and continual service improvement), you may be able to reduce headcount while maintaining satisfaction levels. One of our clients was forced to reduce their Service Desk headcount by 50% but was able to increase customer satisfaction at the same time. Their average resolution times blew right out, but customers were still happy because their expectations were being carefully managed and because the remaining Service Desk staff were focused on delivering a great customer experience. If you’re not happy with the way you collect and analyse customer feedback, our CIOPulse product might be for you.

Good luck and do let me know if you’ve found a good way to determine your Service Desk headcount.

This Post Has 3 Comments
  1. That was an interesting article. I would like to know your opinion about Managed Services vs in-house service desk. Which is option is beneficial as per you?

    Thank you,

    1. Hi Milind. I am a firm believer in keeping the Service Desk in-house. The Service Desk is the ‘face of IT’ and IT’s reputation is heavily dependent on how the quality of service from the Service Desk is perceived. It’s therefore important that the Service Desk be heavily focused on customer satisfaction. In my experience, it is easier to make this happen with a team you manage than one managed by an MSP

Leave a Reply

Your email address will not be published. Required fields are marked *