This document discusses shortcomings of traditional service level agreements (SLAs) and proposes an alternative approach using performance decrements. It notes that while SLAs are needed, they are often disliked due to unrealistic terms and paltry penalties. The document then recommends focusing SLAs on major outages instead of minor issues by using bonuses for exceeding targets and penalties for missing critical thresholds. It acknowledges challenges in implementing this approach but provides examples of how decrements could work in practice and lessons that could improve more traditional SLAs.
2. Topics
Why We Need SLAs
Why We Hate SLAs
Breaking the SLA Mold
Decrements in the Real World
An Action Plan for SLAs
2
3. We need SLAs (or something
like them)
Many applications need
performance metrics
Customer experience can be
impacted
Internal stakeholder issues
Getting what you pay for
Get carrier to put skin in the game 3
4. But we also hate SLAs
The network is the network
Stock SLAs are a joke
Inverse credit/performance
relationship
Carve-outs/restrictions
The need to jump through hoops
for paltry recoveries
The vendor doesn’t feel your pain 4
5. So what do we do?
Give up
Get redundant vendors for
everything
Improve SLAs on the margins
Change the paradigm
5
6. Changing the SLA paradigm:
performance decrements
Goal: forget the small stuff; focus
on the BIG problems
Move past pure credits and find
other ways to incent performance
What about bonuses for exceeding
numbers?
6
7. Challenges of decrements
Contract integration
Selling it to the vendor – you need a
certain size of spend/complexity
Pricing considerations
Achieving the right mix of
• points,
• penalties &
• pain
7
9. Decrement case study
What happened
What worked
What needed refinement
Lessons learned
9
10. Can we port any of these
lessons to non-decrement
SLAs?
Yes (to some extent)
• Reporting
• Account team support/escalation
• Periodic SLA reviews
Breaking the rigidity that comes
with focus on the network alone
10