The service level agreement (SLA) model has always existed in the world of IT in some form, as a promise to guarantee a level of service.
For a long time, SLAs have offered very traditional approaches to measure what is perceived as quality of service by IT organizations. These traditional metrics are based on IT infrastructure and applications, such as “server up/down” or response times of an IT software application. When technology was less complex these metrics were fit for purpose; but what about today?
The historical influence of SLAs
IT and IT service management (ITSM) have always been highly influenced by SLAs, influencing behaviours, prioritizations of resources and steerage of relationships. In my opinion, SLAs for the most part, have managed to create a wholly negative culture between IT organizations and service providers.
It could be said that the construct of SLAs is, fundamentally, the reason IT departments are not perceived as innovative and strategic. IT organizations are often seen by the business as underperforming, disconnected from the needs of the business and simply a “commodity” rather than a partner. One of the major factors is that IT continues to design and report on metrics that have little to no value and do not demonstrate how IT is contributing to an organization’s business outcomes.
“Watermelon reporting” is a common phrase often attributed to a service provider’s performance reporting. Typically, these SLA reports depict that the service provider has adhered to the agreed service levels and met all contractual service level targets. It looks “green” on the outside, just like a watermelon. However, the level of service perceived by the business does not reflect the “green” status reported (it might actually be “red”, like the inside of a watermelon); and this is regularly a source of annoyance to the rest of the organization.
For example, a typical SLA model might state that there can be no more than four priority 2 incidents within an agreed measurement window. Surely, once that target is met, the service provider is now motivated to focus on another customer’s priority 2 targets. Conversely, service providers are typically reluctant to agree to binding service levels for priority 3 and 4 incidents; these are typically agreed as “best effects” to restore services with KPIs (which attract no financial penalties). The result is that priority 3 and 4 issues take forever to get fixed and become the bugbear of IT users!
Unfortunately, IT organizations and ITSM frameworks continue to use this outdated SLA model because it’s what they are comfortable with. In my experience, service level agreements are not typically negotiated or fully consulted with representatives from the business (the people that actually use the IT service); negotiations with service providers and the design of service levels is usually a procurement function.
Let’s look at availability SLAs: I have often heard procurement folk make proud statements about negotiating an agreement of “99.8% availability” with a service provider. But what does this actually mean? Suppliers measure availability differently and its relative meaning needs to be closely examined. In the past, it meant the server was up; today, there are so many components, integrations and different suppliers enabling a business process that the meaning has changed irrevocably.
The SLA and combining business/IT objectives
Today, a more meaningful use of SLAs is about measuring business processes mapped to business outcomes. This is in line with the ITIL principle “Focus on value” where outcomes, not outputs, are what matters. Measuring business outcomes also takes serviced consumption into consideration, not only service delivery. This is what ITIL describes as the “co-creation of value through service relationship” in which “customers are an essential element in the process of creating value”.
Taking the example of an airline, it’s now possible to report the availability of business processes, for example, check in passengers, provide the correct meals for special dietary requirements and allocate seats on a flight. This certainly sees a far more customer focused measurement than the previous SLA that measures whether the system overall is available/unavailable.
To arrive at this understanding, it’s necessary to decouple the service and look at what you’re measuring. It really is a “light bulb moment” when you adopt a design approach to providing a service and measuring its delivery. In the airline business, for example, you recognize that the real value of an SLA is enabling customers to complete their journey. Who cares whether the IT system is “up” for 99.8% ?
Similarly, solving the problem of duplicate seating for airlines (where two passengers are booked into the same seat) is measured better by a reduction in passenger complaints rather than counting fewer incidents. The point is to look at the entire customer experience and design SLAs around that instead.
The state of the SLA today
An honest assessment of suppliers’ current SLAs would suggest that a large percentage simply do not make sense.
Most importantly, when questioned, a supplier finds it difficult to articulate what availability is and how it is measured.
Let’s look at another real-life airline example: the common function of printing bag tags. The system was printing incorrect information on the bag tags resulting in major issues including flight delays and poor customer experience. However, when the supplier reported service performance for the month, they had met all SLAs. Let’s distil this: the system was available – it was printing the bag tags – so it met the agreed availability target. Dare I say, this is where “watermelon reporting” was at play.
There is a great opportunity to really evolve and design innovative SLAs using human-centered design that always focuses on the customer experience. One approach is to design service levels that are based on an organization’s business outcomes.
Using an airline again to provide examples of business outcomes:
- An accurate and optimized roster is available to manage crew operations
- Minimal manual labour spent entering regulatory data at airports
- Complete aircraft turnarounds to meet flight schedules and on-time performance.
That said, it would be rare that a single supplier would accept these types of agreements, both from a financial and accountability viewpoint; typically, no single supplier has complete ownership of the outcome.
One solution to meet this challenge is to design “outcome-based agreements”. In another airline example, lost baggage is the highest category of complaints for airlines. The management of baggage involves many service providers and components: providing the destination data, printing the bag tags, scanning and tracking bags. Collectively, all suppliers contribute to an airline’s business outcome: customers’ baggage arrives on time, at the correct destination in the correct priority.
The design of outcome-based service level agreements – agreed with each supplier – would determine how each supplier contributes to baggage management for the airline and allow the organization to report a holistic view to its stakeholders. Imagine, IT being able to report a percentage decrease in lost baggage for a given month (versus 99.8% availability)!
This approach could apply to any industry and what organizations need to rethink and design are SLAs that are sensible, valuable and move away from IT metrics. Re-designed SLAs need to demonstrate how IT organizations contribute to business outcomes – which, ultimately, is all about customer experience.
ITIL: the SLA today and tomorrow
ITIL 4 talks about facilitating value co-creation via a service value system: different components and activities working together to facilitate value creation through IT-enabled services.
My contribution to ITIL 4 is about humanizing the guidance. Why is this important? There are different types of end users including front-line staff and operational people and they all have different motivations, limitations and environments.
Applying design thinking to ITIL means it’s more about understanding human behaviour of users. In my experience, ITIL process owners often make erroneous assumptions about human behaviour; springing from a poor understanding of users’ environments, motivations and their inclination (or disinclination) to call the Service Desk.
ITIL 4 emphasizes the importance of collaboration, transparency, automating where possible and working holistically.
The future of the SLA
The dream SLA model of the future is an outcome-based agreement by which external or internal service providers understand their role and how they collectively contribute to an organizations’ business outcomes. This approach changes the model from a silo-based, “finger pointing” culture to a more proactive and strategic partnership between IT, their service providers and the business which IT is there to serve.
If this new approach is not adopted, IT will continue to address the same issues, suffer from the same business perceptions and generate reports that offer very little use to anybody. While some will argue that historical IT metrics still have a place, I contend they should live in the appendix of the report, not on the front page.
Effective IT organizations must change how they measure in order to be innovative and show the organizations they support how much they’re contributing to business outcomes.
That means IT needs to align metrics to outcomes such as an increase in customer satisfaction, contribution to revenue growth and brand awareness – this is the stuff for the front page.
For further information about ITIL 4, visit our ITIL Update page.
Read more posts in this series
Combining different IT and ITSM frameworks for business benefit
Business value from IT goods and services in a digital world
The ITIL update in a world of digital and service transformation
Best practice in IT, ITSM and the ITIL update