Service level agreements (SLAs) are the heart and soul of many outsourcing contracts. They define what the provider must deliver and their penalties for failure, in anything from application uptime to the time required to solve a customer’s problem on a help line.
But at least as currently defined, SLAs often fall short of detecting (and, more importantly, correcting) problems quickly. That was the message at the recent SIG Spring Summit from Senior Corporate Counsel Richard English of Ingram Micro and Shaalu Mehra of Sheppard Mullin Richter & Hampton, who helps the electronic distributor negotiate outsourcing deals.
SLAs fall down, said Mehra, because they don’t change with the customer’s requirements, aren’t defined precisely enough, and often aren’t structured to do a root-cause analysis of the root problem behind multiple failures.
“I love SLAs,” Mehra said in a session on “Best Practices for Ensuring Quality of Service in Multinational Outsourcing Engagements.” However, he continued, they are limited because they are just one “data point” measuring a provider’s performance.
While SLAs are the subject of intense negotiations at the start of engagement, he says, they may not be based on the right metrics to measure the effectiveness of the outsourced service for the customer. In addition, he said, SLAs “can be undermined by even minor changes” to the processes or systems they measure, and are often not updated often enough.
Another factor that limits their usefulness is “single incident limitation, (which makes) root cause analysis subject to an agreement of the parties,” said Mehra. Understanding and correcting the reasons for past failures can also be hindered by what English called a “statute of limitations” requested by vendors on how long a customer can ask for a service credit after a failure.
Both strongly suggested using a common approach to SLAs and other terms with all outsourcing providers, regardless of their location. “While that company may be India-based, and might be doing work for us in Asia, or might be doing work for us in Latin America…we don’t care,” said English. “We’re going to build in one global SLA.”
Two areas where the pair said specialized SLAs might make sense were to measure English language fluency and attrition. While Mehra agreed that “fluency” is subjective, he said it could be measured through a sampling of calls or surveys of whichever end users were being served.
Mehra said such SLAs often measure attrition on a rolling 12-month basis, and is an area where definitions (such as whether promotions, reductions in force or departures for personal reasons count as attrition) are often the cause of hard bargaining. “However reasonable the vendor’s concerns may sound, at some point, we have to draw a line” about where and how the customer will be protected from excessive turnover, he said.
In addition to SLAs, Mehra recommended regular payments based on the achievement of milestones, as well as periodic payments with provisions for holdbacks as a penalty to the provider for failures in delivery.