Key Service Level Management Concepts Explained

In the previous article, we explained how to:

• Build value-driven service levels by establishing service level targets aligned with business needs, target audiences, services scope, and installed capabilities
• Use data (historic performance, market standards) to set specific target values
• Prevent value erosion in service levels

Our latest blog explains the importance of service level management, key concepts used to manage service levels with third-party vendor relationships, and how these can be effectively used to incentivize third-party vendor behavior.

To keep it simple, many people understand service level management (SLM) as the practice to monitor, report, and review the performance of services being delivered, focused on ensuring that the services perform according to agreed service targets. With this perspective in mind, when service levels are poorly designed (not measuring what is important) or service targets are easily met on a regular basis, SLM devolves into a mere administrative exercise.

When designing SLM processes, focus on how to deliver enhanced value, not only through the selection of the right metrics, but also through the actions triggered by the data provided. It isn’t about collecting data, but rather translating data into action.

SLM processes should help organizations:

• Incentivize the right behaviors from service delivery providers through service credits or incentive fees
• Identify and address issues and risks through the identification of trends, root cause analysis, and the implementation of preventive and/or corrective actions
• Determine continuous improvement opportunities, and monitor and assess their impact upon implementation

Service levels are not only used for performance management, but also as a way for third-party vendors to demonstrate their commitment to your organization through risk sharing and prevention, and continuous improvement.

A key aspect of SLM is the need to incentivize the right behavior from a third-party vendor. The nature of the services and the relationship (e.g., when risks are shared) will determine the mechanism(s) to be used. This can be done through any or both of the following approaches:

At-risk amount is defined as the maximum amount that the third-party vendor is willing to pay as a penalty for not meeting the defined service level targets during a payment period (usually monthly). The at-risk amount is usually a percentage of the monthly bill for the services being delivered. For example, a vendor may agree to put 10% of its monthly revenues at risk for failing to meet agreed service level targets.

This number can be linked to the level of confidence that the third-party vendor has in the quality of the products or services delivered. Having a relatively low amount reduces the third-party risks, does not provide an appropriate incentive, and erodes value for the organization. The amount needs to be fair for both parties, and adequately compensate for portion of or all the cost associated with the management of service levels and the business impact caused by inadequate performance.

To determine the right percentage, organizations must consider several factors:

1. The basis of calculation: Are the total monthly fees at risk or only monthly fees related to the service level target in question?
2. The monthly billing amount: Are we talking about enough money to make the vendor financially concerned about incurring a service credit?
3. The cost to manage service levels: Is it costing more just to measure the services than the vendor is putting at risk?
4. The potential business impact associated to the services provided: Are we being fairly compensated for the lack of service quality?

Pool allocation is defined as the percentage of the at-risk amount that can be allocated across all critical service levels (CSLs). The pool allocation enables organizations to increase the impact of inadequate performance for specific CSLs to better align them with the business impact.

Many third-party vendors will propose an allocation of 100% of the at-risk amount across all defined CSLs. In some instances, when the number of CSLs is low, or there is a low impact to the business, this can be acceptable. When organizations need to allocate the at-risk amount across a high number of CSLs, or have CSLs that, due to the business impact, need a higher allocation percentage (even up to 100% of the at-risk amount), it is necessary to increase the allocation significantly (e.g., 250% to 300%) to prevent value erosion.

For example, if I have a 10% amount at risk on a $1 million per month contract, that’s a potential $100,000 service credit—it seems we have the vendor’s attention! However, if I need to measure 20 service levels and spread the amount at risk evenly, I will only receive $5,000 for a missed service level—not a huge disincentive for the vendor. Perversely, it may be cheaper for a vendor to not invest in improving a service and paying an occasional $5,000 service credit —not the outcome you want as a client.

It is important to understand that the at-risk amount remains as a cap that can’t be exceeded by having a pool allocation. It only enables a single CSL or a combination of CSLs to reach the cap faster.

To determine the right percentage, you must consider the number of CSLs defined and the potential business impact associated to each CSL.

An individual service level cap is defined as the maximum allocation percentage of the at-risk amount that can be allocated to a single CSL. This cap is used by third-party vendors to cap their exposure or risk in case any of the individual CSLs are missed.

The percentage allocated to each individual CSL should be linked to the potential business impact and not subject to a cap. There may be instances, given the importance of a specific CSL, where such CSLs can trigger a service credit of up to 100% of the at-risk amount.

If used, the cap should not limit the sum of all assigned allocation percentages to reach the defined pool allocation percentage.

Multiplier or termination events are linked to deteriorating performance and are useful tools to help incentivize action on consistently missing CSLs. Although uncommon, organizations need to prevent a situation where the third-party vendor may feel more comfortable paying the service credits associated to a missed CSL instead of investing to fix the problem (as described in the earlier example).

The terms “Multipliers” or “Termination Event” can be used to increase the service credits or terminate the contract for cause when one or more CSLs are:

• Missed for consecutive months
• Missed frequently within several measurement periods (e.g., three out of six consecutive months
• Performing well below the established service level target
• Missed in the same measurement period

There are other aspects that organizations need to consider when defining SLM processes and the associated metrics. This includes the number of CSLs to be used, the measurement/reporting periods, and the methodology/tools that will be used for collecting and reporting the data.

A.Number of Service Levels

SLM needs to deliver value to the organization by focusing on those metrics that deliver the highest value. All others can be measured and should be considered as key performance indicators (KPI).

In general, do not overmeasure; if you are not doing something with metrics, or if it delivers little or no value, don’t use it. Carefully assess the value of each proposed metric as you go through the definition process.

Some of the main reasons to maintain a reduced number of CSLs are to:

• Increase focus on CSLs linked to business value
• Reduce the complexity and cost of collecting data, reporting, and managing CSLs
• Enable a higher percentage of the pool allocation to each CSL, increasing the impact/incentive for the third-party vendor

B.Measurement/Reporting Periods

CSLs can be classified based on the recurrency of the events linked to that CSL. The type of CSL will determine several special considerations related to the measurement and reporting periods:

C.Data Collection Methods/Tools

The rule when defining the CSLs, to the extent possible, is to keep them simple. Data collection and reporting can be expensive. When defining the CSLs, think about the process to collect, process, and report the data. Organizations need to ensure that the data provided is accurate and enables the right decision-making. In cases where you are relying on your vendor to collect and report the data, then you need to consider the level of transparency and/or your ability to inspect the raw data itself to ensure CSL performance calculations are valid.

Third-party vendors often propose elements to be included in service level calculation and management. In cases like this, you as a client need to be watchful because some can erode the value of the SLM process:

A.Earn-Back

Earn-back implies that if the third-party vendor paid service credits associated with the poor performance of their service, they are entitled to recover those credits if they overperform in subsequent periods.

While the argument seems fair, it does not address the following:

1. CSL targets are minimum performance expectations. They assume that the organization will suffer damages below this level, but there is no guarantee of additional value for any performance above the established CSL targets.
2. The organization will not recover any business impacts or costs associated to the poor performance of services, even if the third-party vendor overperforms in subsequent months.
3. There is no way to define the right overperformance metric. As mentioned before, CSLs are minimum performance expectations, but the actual expectation is that the third-party vendors will consistently deliver above CSL targets.

B.Multiple CSL Failures by a Single Event

Third-party vendors require that, when multiple CSL targets are missed due to a single event, only one of those CSLs should be considered as missed. This is intuitively logical, as the law recognizes and prevents “double jeopardy” (this is such a fundamental concept, it’s even enshrined in the Fifth Amendment to the US Constitution). Therefore, to the extent it is possible, CSLs must be as independent and exclusive as possible.

Organizations can define CSLs tied to end-to-end services that incorporate all individual components into a single metric (e.g., end-to-end application availability). In other cases, organizations can identify and call out certain CSLs that should be excluded or treated as an exception, given that multiple business groups may have been impacted by different service metrics linked to a single event.

In this article, we discussed the importance of building SLM processes that incentivize the right behavior from third-party vendors, including key concepts that are used to properly manage the services and prevent value erosion. Third-party vendors need to show confidence and commitment in the quality of the products or services they deliver, assume accountability, and be incentivized to take action to prevent and/or correct any issues.

In the next and final article, we will do a deep dive on the type of actions that can be triggered by a well-defined SLM process and how to deal with third-party vendor concerns.