I recently had a discussion with one of the respected global company that specializes in providing consulting around outsourcing and managing Service Level Agreements. My friend on the other side of the phone passionately tried to convince me why it is important to create SLAs that can tie down every aspect of the service that the outsourced service provider will deliver now or in the future. He cited many examples of how his company helped many customers “win”.
In another setting a debate was set off between CIOs on how they ensure that their service providers deliver what they promise consistently that meets the promise to the business. For more than a year, one of them has been unsuccessfully trying to get a bunch of vendors to come to the table for a discussion on creating effective SLAs. Not that the vendors are shy of the subject but collectively at the same table with multiple CIOs is not a viable proposition.
Service levels matter to everyone, the customer, the provider and the end consumer of the customer; I do not believe that deficiency of service is due to wilful behaviour or mal intent. The exception to this may be in monopolistic scenarios where no incentive exists. When it is relatively easy to switch services or move business to competition, efforts are indeed put in by the provider, the end results may however not be aligned to expectations.
The reasons why SLAs fail could be many ranging from ambiguous definition of service, staff involved in execution not being aware of quality of service expected, lack of skills on the ground, unrealistic expectations, or force majeure conditions to name a few. Irrespective of the reasons, when things do go wrong, contracts come out of the closet again to review the penalties that can be levied or avoided depending on the frame of reference. My belief is that “if-then” motivation will not deliver world class service; i.e. if SLA is met you get paid, if you better the SLA, you collect a bonus, whereas if the SLA is breached, there is a penalty.
SLAs are typically calculated on statistical data which fails to recognize business impact when the service is deficient. Creating complex SLAs that factor in all types of exception conditions makes it readable and enforceable only by lawyers and not CIOs. A SLA should illustrate the intent of partnership between the two (or more) parties. Incremental innovation or improvements are expected as much as occasional failures that could be for any of the reasons listed above. Both parties need to work together towards ensuring that they understand the root causes and work towards prevention of repeated adverse impact.
Unfortunately such behaviour is rarely seen and everyone invests significant resources towards the scripting of a document that covers all bases. End result is that the parties involved split hairs with irrational discussions thereby leaving the spirit of partnership aside. Most successful relationships are based on simple few page documents that capture the intent with the managements investing time in frequent reviews not just when things go wrong, but when they are working too.
Over the years it has been a difficult journey on this path, but it has been worth the effort. The big companies (customer as well as provider) have however yet to learn.
In another setting a debate was set off between CIOs on how they ensure that their service providers deliver what they promise consistently that meets the promise to the business. For more than a year, one of them has been unsuccessfully trying to get a bunch of vendors to come to the table for a discussion on creating effective SLAs. Not that the vendors are shy of the subject but collectively at the same table with multiple CIOs is not a viable proposition.
Service levels matter to everyone, the customer, the provider and the end consumer of the customer; I do not believe that deficiency of service is due to wilful behaviour or mal intent. The exception to this may be in monopolistic scenarios where no incentive exists. When it is relatively easy to switch services or move business to competition, efforts are indeed put in by the provider, the end results may however not be aligned to expectations.
The reasons why SLAs fail could be many ranging from ambiguous definition of service, staff involved in execution not being aware of quality of service expected, lack of skills on the ground, unrealistic expectations, or force majeure conditions to name a few. Irrespective of the reasons, when things do go wrong, contracts come out of the closet again to review the penalties that can be levied or avoided depending on the frame of reference. My belief is that “if-then” motivation will not deliver world class service; i.e. if SLA is met you get paid, if you better the SLA, you collect a bonus, whereas if the SLA is breached, there is a penalty.
SLAs are typically calculated on statistical data which fails to recognize business impact when the service is deficient. Creating complex SLAs that factor in all types of exception conditions makes it readable and enforceable only by lawyers and not CIOs. A SLA should illustrate the intent of partnership between the two (or more) parties. Incremental innovation or improvements are expected as much as occasional failures that could be for any of the reasons listed above. Both parties need to work together towards ensuring that they understand the root causes and work towards prevention of repeated adverse impact.
Unfortunately such behaviour is rarely seen and everyone invests significant resources towards the scripting of a document that covers all bases. End result is that the parties involved split hairs with irrational discussions thereby leaving the spirit of partnership aside. Most successful relationships are based on simple few page documents that capture the intent with the managements investing time in frequent reviews not just when things go wrong, but when they are working too.
Over the years it has been a difficult journey on this path, but it has been worth the effort. The big companies (customer as well as provider) have however yet to learn.
No comments:
Post a Comment