As law firms change in structure and size, often expanding globally, their IT support model also has to evolve. With a great deal of growth expected to come from overseas countries, particularly from Asia, some alternative thinking is required to accommodate the scaling of IT support services to meet the increased, non-local, demand.
Current IT support delivery models are likely to be centred around a large centralised Service Desk providing all in-hours support and which is provided by an in-house team. Out-of-hours and global support is then provided by any number of different configurations, often including complex rota systems utilising staff and/or contractors, or an externally provided shared service (perhaps even a combination of the two). Such 24×7, or ‘follow-the-sun’, solutions do work and may be reasonably inexpensive but are frequently difficult to manage and do not provide a consistent level of good service across all locations.
Additionally, we’ve also recently witnessed a change in demand from end-users based in what have traditionally been outpost office locations. These end-users, aware that they’re based in locations of new growth, now expect the same level of service as their counterparts in the more established offices. This means receiving the same quality, response and availability of IT support services. Arguably, these demands are reasonable and there should no longer be a difference in the IT support received, regardless of location.
To summarise – with growth predicted to come from abroad, and with a demand for a consistent delivery of services, regardless of office location, a revised approach to the provision of IT support services is needed.
An alternative concept of providing IT services may be to work to the ‘troughs’ in demand and not the ‘peaks’. Let’s explain what is meant by this.
The bulk of an IT support service, perhaps 70 to 80% of it, is provided during a core period and is utterly predictable. This portion of the service represents the ‘trough’ in demand, i.e. it is unmoving, consistent and therefore easy to plan a team around. The ‘peaks’ in demand are portions that are prone to variations, e.g. the magnitude of spikes in demand, the volume of out-of-hours support activity, or the support demands from overseas offices. It’s the smaller and less predictable portion of the service (the 20 to 30%) that consumes a disproportionate amount of management time, whilst still resulting in an imperfect service.
The trough and peak concept essentially shifts the main focus of in-house service provision to the larger and unchanging part of the service, where service excellence is the goal. A more scalable approach is adopted for the peaks which strive to meet the same levels of excellence as the main portion of the service whilst easily being able to deal with the variations, however subtle, in demand.
Traditionally, an IT support service must align its own capacity model with its demand curve, i.e. when demand for support begins to build on a weekday morning, then the Service Desk opens. As the demand curve increases during the morning (to its peak at around 11am) so does the number of support analysts that are available. Then, later in the afternoon, as demand falls, so does the staffing on the Service Desk until, ultimately, it closes. At this point, an out-of-hours function takes over until the cycle starts again in the morning. With out-of-hours support activity, a measured amount of resource is available that can manage the anticipated out-of-hours, and/or global support requirements, until the main Service Desk reopens.
The single biggest issue with the traditional method is that resourcing for both the in-hours support needs, and the out-of-hours/global demands, is based upon meeting the peaks in demand. If the service needs to be extended, all that can be done is to add analysts to the capacity model which will see significant step change in cost.
The following line graph shows a fairly typical weekday demand curve (i.e. ticket logging activity by hour of the day). Demand builds in the morning and tails off at the end of the day. The main peak is mid-morning with a smaller peak in the middle of the afternoon. In this example, the bars represent a capacity model based on 7 analysts working 7.5-hour days, with staggered starts, and an hour for lunch, with Service Desk opening hours of 7am to 6pm.
On the face of it, it could be said the capacity model is a good fit for the demand curve. But on closer inspection, the capacity model has failings:
1) Demand begins to build before the Service Desk opens. These calls may be picked up by the out-of-hours service, but the people manning that service will soon be finishing their shift, so they may just ‘log and flog’ so that the support can be picked up by the in-hours team;
2) Resource does climb throughout the morning in line with the demand curve – however, the maximum available capacity isn’t quite enough to meet demand at the busiest time of the day. This may well lead to an increase in abandoned calls at this time;
3) Then, due to staggered lunch breaks, resource again doesn’t quite keep in step with demand;
4) The analysts’ shifts end slightly ahead of demand, and finally:
5) The Service Desk closes, switching to the out-of-hours service, just before the day’s demand has fully settled down.
By working to the peaks in demand, our well-considered capacity model doesn’t quite fit demand (and nor will it ever), and if we need to scale-up in line with business growth, all that can be done is to add extra analysts and attempt to further match the capacity model, as best as one can, with the demand curve.
Managing resource for an out-of-hours service has similar challenges. A capacity model must be derived that meets maximum demand. If the out-of-hours service is based on staff, equipped with phones and working to a rota, such a rota can be very difficult and onerous to manage. If the service needs to be expanded, all that can be done is to add resource to where the gaps are emerging, even if it means that the resources will be underutilised.
If, however, an in-house core service is delivered to the troughs (which will still be the greater part of the service), and a suitable high-performing shared service is identified which can handle the peaks, i.e. the lower-volume out-of-hours activity, peak time overflow and variations in throughput, then an improved and scalable service model will be achieved.
The following diagram shows the same demand curve but with a reduced in-house capacity model.
In this example, the box represents 5 analysts, as opposed to 7 in the earlier diagram (a 28.6% reduction in headcount). This team would still manage 71.4% of the in-hours call volumes and can spend its time focusing on delivering an excellent service. The remaining support, i.e. the peaks, the increasing and decreasing demands at the start and end of the normal working day, and the out-of-hours piece, can be handled by a shared service which integrates with the main in-house service, but which can manage variations in demand with ease, scale up or down when required, and may demonstrate cost savings.
To use a term that I like to use, this model allows you to concentrate on the steak and not the peas. Instead of working very hard to manage the smaller percentage of support activity at the edge of your service, focus on the bulk in the middle, and utilise the availability of a well suited outsourced service to manage the smaller, less predictable volume. Savings can be made, the all-round service improves, you have a scalable model that’s less onerous to deliver and your customers (all of them) are happier.
It’s worth finishing with some guidance on how to transition from the current model to a new one. The answer is found in the data contained within your IT Service Management tool. It is essential that your existing service is fully profiled, from ticket origin through to logging, first-time-fix, escalation, and finishing at resolution. An effective data analysis process needs to be conducted on your ticket management data so that your full service, and how it is used, is fully understood. This kind of analysis is typically above-and-beyond the standard reports you might get from your ITSM tool, but instead is a data-mining exercise which performs a distillation of your service, against which a new model can be defined. With such an analysis behind you, the decision of how much to keep in-house, versus what to outsource, is a reasonably straightforward piece of cost analysis. Whilst adopting the trough and peak model may seem too greater a step to make, the required due-diligence to see if it’s viable won’t impact the delivery of your services at all. So, at the very least, it would seem to be a prudent move to investigate it as an option.
Jon Reeve, Principal Consultant
This article appears on ITSM Portal: http://bit.ly/LzkyGG