Optimizing your SLA or managing a watermelon

An important part of any sourcing engagement is the Service Level Agreement. The SLA lists the agreed (Key) Performance Indicators and the norm for each to which a vendor should comply. In some cases penalties apply when service levels fall short of the agreed norm. Despite the serious effort usually spent during negotiations into agreeing the service levels and associated reporting the result may still be unsatisfactory. A quite common occurrence in day to day service level reporting is the so-called watermelon: green on the outside but red within.

A watermelon service level report shows all the KPI’s within the agreed norm (green) while the underlying services at the same time are perceived by the relevant business functions and or by the end-users as less-than-satisfactory (red). Obviously the client will feel the need to make changes in this situation but also from a vendor point of view this type of relationship is not sustainable. Therefore parties will need to address this jointly and find a way to agree on service levels and associated norms that better support the business needs.

The most important requirement to any service level is that the measurement is fact based and both parties should have access to the relevant data on which the reporting is based. A number or a percentage that can be calculated unambiguously avoids a lot of unnecessary discussion about service level reports.
Also the customer satisfaction surveys need to be objectified as much as possible. Obviously there is some subjectivity in determining satisfaction levels but by making questionnaires multiple choice, by using consistent sets of questions in subsequent surveys and by including a significant number of respondents from the relevant population, the basic statistical errors can be avoided.

Optimizing the SLA can be achieved by reconciling the differences between reported service levels and perceived customer satisfaction. In any case where the norm is met while business functions or end-users rate the service poorly (or the other way around!), specific action is required. There is no one-size-fits-all solution for this but a good starting point is to align business value as close as possible with ICT service levels e.g. availability of a sales platform during peak hours, optimized response times for key business users etc. The challenge is to be creative here without hugely increasing the complexity either in measurement or in reporting requirements.

Obviously the watermelons need careful attention from both parties but also in cases where customer satisfaction exceeds the reported services levels, further analysis is highly desirable. Service level reports should be like avocado’s: green all the way to the core.