Question: If you have a client SLA of 3 days, how long do you have to complete that task?
First, a little background:
SLA stands for “Service Level Agreement” and is how clients measure the performance of their providers. Generally, an SLA would read something like this:
“95% of widgets are to be built within 3 days.”
This leaves 5% for exceptions, but the general rule is that you need to get nearly all of your widgets built within 3 days.
So, how long is 3 days?
That might seem like an easy or even silly question. Everyone knows how long 3 days is. Ok. Now how do you measure it?
Let’s keep things simple and assume our client runs on a 24-hour clock and counts weekends and holidays. This means they haven’t specified business hours and the clock is running 24×7.
For our example, let’s say we received an order for a widget at 10:00 AM on Monday, April 1st, 2019.
Exactly how long do we have to complete our task and successfully fulfill our 3-day SLA?
My math, based on counting on my fingers, says we have until exactly 10:00 AM on Thursday, April 4th to complete the task within 3 days. If we were to measure the start time to the end time, it would measure as 3.0 days. (note the decimal, it’ll be important soon.)
Why am I writing this? Isn’t this obviously the case? Well, no.
In my years in a business intelligence role, I’ve had a few people from the business tell me this is wrong. That, the measurement for success continues on to 3.999… days. This would effectively give them until nearly 10:00 AM on Friday to be successful on a 3-day SLA that began on Monday. In the example below, I believe we’ve failed a 3-day SLA once you cross past Stop A and these people argue that they have up until Stop B.
In these cases, I’ve expressed my disagreement, but occasionally, a business leader persists with their argument and I really need to convince them. I’ve drawn number lines, asked a million hypotheticals, counted on my fingers, even used wooden blocks, but some still won’t come over to my way of thinking. I’ve now had this conversation across enough companies and enough different industries to know it’s not an isolated incident. There is a segment of people who believe that 3.999 days is a successful value for a 3-day SLA and I’m entirely unable to convince them otherwise. This is to the point where I’m questioning my own methods and considering that I might be wrong. Therefore, I’m asking the internet to convince me. Can anyone make an argument that 3.999 days should equal success on a 3-day SLA?
Let me note that this isn’t limited specifically to 3-day SLA’s. The same argument has been made for any number of days, including an SLA phrased as 24 hours. There was even one where the SLA was 4 hours and I was told we should count it as successful if we were done within 4.999… hours.
What are your thoughts? Can you convince me that 3.999 is less than 3? How would you handle this situation?
If you agree with my point of view, how could I better express this concept to business leaders so they understand the “right” way to calculate their SLA’s?
I will not confirm or deny any calculations made at current or previous employers. The question was asked, not necessarily implemented.