Five to 10 years ago (depending upon the sector), most service operations ran on the basis of "best endeavours": service level agreements (SLAs) or guarantees were to be avoided.  Service was still considered as after-sales support and about repairing faulty equipment at a time when the price of service was almost cost-plus (in other words a percentage of the sale price of the equipment). 

Contracts were loosely based and the pressure on service levels relatively relaxed; circumstances that many sectors of the service industry still enjoy.  Many companies now have as much as 80% of their client base under contract, with strict SLAs in place.  Initially (out of ignorance), SLAs had little importance placed in them and were often either arbitrary, or represented what it was possible or easy to measure, consequently they were either too generous or too
draconian, but certainly not very representative of the true situation. 

Research indicates that over the last five years, customers and suppliers alike are taking SLAs much more seriously, and customers are increasingly invoking the penalties written into SLAs.  Such customer attitudes, combined with falling prices, have caused many problems for suppliers, including being replaced: yet few customers have been abandoned even when no margin is being achieved from their custom.

There is a clear need to measure, but measures need to be much more specific and detailed, and should consider any trade-offs that may be necessary.   An example is shown diagrammatically below. 

The figure illustrates that most service operations now measure engineer travel time and on-site time and offer an average per visit; research shows that the ratio between site and travel time seems to be about 25% to 30%, but what this measure does not reveal is any insight as to how the travel time is made up nor how to reduce it. 

In one case study the exercise was to look into the detail that made up the total number to understand what really made up the time allocated as travel time.  The revelations included that there was no formal agreement among the staff or management as to what truly constituted travel time, and across a number of engineers travel time could be booked in a number of ways.  The time taken to acquire a part that was available either at stores or with another engineer, could be booked to travel or site-time depending on which aspect the engineer felt would balance his day.  In research over the last 10 years, this measure has remained an almost constant 30% because it would seem the industry expects this result and appears complacent about it. 

This example serves to illustrate the necessity for detailed accurate measurement, not as an optional extra, and it also has to be understood that the mathematics behind this level of calculation is becoming too complex to be done on a spreadsheet or in an Access database. A software tool is therefore required to recommend the necessary actions. 

If the operation is to be effective, the operation measuring SLAs must have detail around the failed SLAs, and able to give chapter and verse as to why the SLA failed and where the fault lay.  This is not intrinsically a problem, but what may be is that customer-driven service levels are vulnerable to finding fault with the customer and, if possible, blaming the customer for failures in SLAs. 

Many service operations measure only the failed SLAs - because they do not possess the capability or have the systems in place to efficiently measure the detail necessary.  There is always a balance to establish between customer satisfaction and performance, but it can be dangerous to drive for customer satisfaction without understanding the cost of doing so.  A review of one service operation revealed that it would be financially better off if a percentage of the SLAs failed and the penalty paid, rather than achieving the SLA and paying the additional costs of emergency parts, shipment and coping with the disruption. 

Investigation into all calls for a particular region, whether achieved within SLA or not, found that a vast number of the visits within the SLA were only just achieved.  One could argue that this could be rated as good performance, however when the detail is explored, the reasons for the call coming close to failing the SLA were owing to limitations in the processes and did not reflect a success through good management.  Improving the basic process in many cases would have produced a much lower cost result and provided capacity to deal with the calls failing SLA, as quite often the close shave would cause another call to fail. 

The nearly-failed SLAs highlighted: procedure problems, lack of process discipline, stock deployment issues, skill issues, call closure etc, and provided significantly more information on the drivers of performance within the operation than looking at the failed SLAs, probably because no one was trying to find an excuse when the SLA was achieved.  (Boasting about a close shave and how no cost was spared to achieve the SLA is not necessarily a success story). 

Service operations set high standards of achieving a set number of calls per day or some similar productivity measure, but fail to assess what may have an impact on that productivity.  As service operations agree to more and more aggressive SLAs in an attempt to win business, there will be a real need to measure the operation to a degree of accuracy appropriate to the situation.  Sometimes the 'stretching SLA' is not obvious, as some contracts gradually creep in scope from the original, (though if the necessary measures are in place this creep will be quickly noticed and dealt with before becoming the standard).

It has become more and more challenging for a business to understand what their operation is capable of when running normally: understanding their capacity as opposed to their performance.  Understanding what the business can do allows it to deal with the fundamental issue of what the real cost of achieving their SLA is. 

A lack of knowledge of capacity is obviously a problem when cutting costs and may be realised to the detriment of the business when SLAs are failed.  In an environment of growth it is an even bigger problem, because quite often the growth will not be gradual but caused by a new contract and if the calculations of resource etc. agreed upon, are based on poor information, the result can be catastrophic at the beginning of a contract.  What tends to happen is that the business initially seems to cope with the additional load and then cracks begin to appear.  If it appears resilient there is a good chance that there is a large amount of capacity available,  however lack of support to the field in releasing calls, or  providing parts will cause the level of failed SLA to grow exponentially.  If these measures are only triggered when a given threshold is breached, then the control becomes more of a knee jerk reaction, resulting in poor management by those not understanding the situation, or how to improve the situation.

The service environment has become much more complex over the last few years; putting the right measures in place can demonstrate significant ROIs that were just not conceived of 5 years ago.  For many businesses these opportunities are currently hidden in the detail where the devil lies!

Effectively managing (measuring) the business is where the difference begins to show, and with the complexity and variation in SLAs (now virtually an individual fit per customer) the need for effective measurement and appropriate management tools is becoming obvious and essential.  As is the need to be able to identify the true value of these tools through more effective and appropriate ROIs highlighting the real trade-offs and providing the necessary understanding of the detail to make effective informed decisions. 

steve.downton@downtonconsulting.com        www.downtonconsulting.com