MoFo Lunch & Learn. Service Level Agreements and Operating Level Agreements. Presented By Alistair Maughan. 13 July 2015

Service Level Agreements and Operating Level Agreements Presented By Alistair Maughan 13 July 2015 © 2015 Morrison & Foerster LLP | All Rights Reserv...
Author: Alexia Palmer
31 downloads 1 Views 773KB Size
Service Level Agreements and Operating Level Agreements Presented By Alistair Maughan 13 July 2015

© 2015 Morrison & Foerster LLP | All Rights Reserved | mofo.com

MoFo Lunch & Learn

Lunch & Learn  2nd Monday of each month  45 minutes via webinar  Unaccredited CPD points  Next session: We will take a break for the month of August. The date for the next Lunch & Learn will be:

 Monday, 14 September 2015

This is MoFo.

2

Today



Questions at the end. Or by email afterwards.

This is MoFo.



Phones are muted to reduce background noise



We’ll unmute at the end

3

Service Levels: Why Are They Necessary?  What’s the meaning of life (or the point of service levels)? Glass Half Empty



Enforce



Punish



Deter



Glass Half Full



Prioritize and Modify Behaviour



Fix problems



Highlight trends



Align Pricing with Level of Service

Don’t assume. Ask (but be ready for an answer you didn’t expect).

This is MoFo.

4

What’s It All About? Whatever the key objectives, Service Levels should NOT be about legal remedies. The top 5 reasons for Service Levels are: Remediation

Clients should be more interested in a service that works than in cash-back for a broken service. Service Levels trigger remedial actions and a get-well plan

To Reflect Reality

Service Levels are there to reflect the user experience. The closer a Service Level scheme is to achieving that goal, the better it will be

Prioritization and Behavior Modification

Not all services are created equal. Service Levels should be there to prioritize the Services Schedule in terms of the value to the client’s business

Baselining

Service Levels should act as a benchmark to measure service improvements

Pricing Fairness

Service Levels can be used as a fair way to tweak the price quickly to reflect service not delivered

This is MoFo.

5

Response to Clients What to do with a customer who wants to use Service Levels to punish? 

There’s no such thing as 100% error-free services



Service Levels weren’t 100% pre-outsourcing, so the deal shouldn’t be about automatic guarantees now



Focus on the cost consequences of high service levels and aggressive legal remedies



Work through escalation and how examples of how underperformance would be dealt with



Focus the customer on the importance of remediation and failure prevention – align any “punishment” element to repeat offences

This is MoFo.

6

What Makes a Good Service Level? A good Service Level is:  capable of objective and reliable measurement  important to the client’s business objectives  prioritized for key business events or touch points to the outside world *  realistic  part of the end-to-end service  specific – not a “goal” or a “target” *

Possibly more for customers to worry about

This is MoFo.

7

Right-Sizing Service Levels Can there be too many Service Levels?

YES



Too many Service Levels leads to over-measuring and underprioritization



Measuring and tracking an excessive number of Service Levels can cause clients to devote significant resources, thus increasing retained costs



While fewer in number, end-to-end Service Levels provide better information regarding the client’s actual business than the generic Service Levels (e.g., application availability is a better measure of operational effectiveness than server availability)



Pay attention to any subjective measures: feed into governance discussions, not specific remedies

This is MoFo.

8

Where to Start? Don’t Drown in Detail  Start at the beginning with the Services Schedule ….  … and then throw at least half of it away!  Not everything that can be measured is worth measuring  Agree what’s really important and create different levels of reporting

This is MoFo.

9

The George Orwell Approach to Service Levels Services

Measured Services

Critical Service Levels Chronic SLA Failures

Statement of Work

This is MoFo.

Service Levels

Service Credits

Termination Right

10

Where to Start? Initial Service Levels 

Initial Service Levels usually set by:   



Pre-contract performance Market standard data – Gartner and similar data sources Typical offerings from service provider (e.g., gold, silver, bronze standards)

Issues    

Lack of client performance tracking pre-contract Effect of Transformation creates uncertainty in steady-state Current environment may not be set up for effective measurement Creating a consistent and verifiable measuring methodology

 To counter these issues, providers will often ask for Service Level relief, or that Service Levels be considered only as key performance indicators (KPIs) for a certain period after measurement is completed

This is MoFo.

11

Service Levels and how to measure them Measurement is the hard part that gets forgotten  Be aware of different types of Service Level   

continuous measurement Service Levels – e.g., system availability event-based Service Levels – e.g., call answering; policy/claim processing sample-based Service Levels – e.g., accuracy of queries dealt with

 Each type has a range of things to consider:    

start/stop points of measurement planned outages duration of measurement – e.g., weekly, monthly, quarterly, annual measurement by hand or automated tool?

 Two-fold approach: Service Levels and Minimum Service Thresholds 

Ratchets and the multiplier effect

This is MoFo.

12

Types of Service Levels Critical vs. Important vs. KPIs   

Critical Service Levels are those that are deemed most important to Service delivery – Subject to Service Level Credits/Breach Important Service Levels, while measured and important, are less critical to Service delivery – Not subject to Service Level Credits/Breach Key Performance Indicators – Measured, not subject to Service Levels or Breach

Critical Service Levels and Important Service Levels can be reclassified (i.e., Criticals can become Importants and Importants can become Critcals). However, KPIs can never be anything but KPIs.

Expected/Target Performance vs. Minimum Performance   

Expected Performance is the level that the provider is EXPECTED to meet every month, but if it fails, generally there is no automatic Service Level Credit Minimum Performance is the critically low level of performance that will trigger a Service Level Credit when it is missed Every Critical Service Level and Important Service Level should have both of these measures

This is MoFo.

13

Continuous Service Level Improvement Premise   

Customers have not optimized service delivery in their outsourced functions Providers bring superior tools, best practices, efficiencies and scale that should result in better performance than the customer was able to achieve The initial Service Levels are based upon historical customer performance, which is non-optimized

Conclusion    

Providers should be able to improve upon the initial Service Levels The improvement should occur year-over-year as provider’s tools and processes are introduced into the environment and provider becomes more familiar with customer’s business Improvements should be formulaic and so not require “continuous negotiation” But, the formula should not be penal, which creates a disincentive to great performance

Typical Continuous Improvement Formula 



Improved Expected Performance = Average of highest 4 months of performance that exceed the former Expected Performance, subject to an increase that is not greater than 10% of the difference between the Existing Expected Performance and perfection Example: Expected Performance was 99.0%; the Improved Performance cannot be higher than 99.1%, even if the calculation of the average of the 4 highest months results in a greater percentage

This is MoFo.

14

Effects of Service Level failure Focus on Practical and Legal effects  What takes priority?  Address service failure escalation issues  Customers should not insist on hitting every Service Level – there is benefit in being flexible  Financial remedies (Service Credits) should be paid timely. Storing until year-end loses the incentive to fix the problem

This is MoFo.

15

Service Level Failures Service Level Failures

Failures to meet Service Levels that results in a Service Level Credit Examples include any of the following combinations: • 1 Minimum Failure • 2 Consecutive Expected Failures of same Service Level • 3 Expected Failures of same Service Level in a 6-month period • 5 Expected Failures of same Service Level in a 12-month period

This is MoFo.

16

Sample Service Credit Triggers  Service Credits occur when there is/are:  1 Failure to meet Minimum Service Levels  2 Failures in a row to meet Expected Service Levels  3 Failures to meet Expected Service Levels in a 12-Month Period EXPECTED SERVICE LEVEL

MINIMUM SERVICE LEVEL

. . . . . . . .. . . . . . . . J

SERVICE LEVEL CREDIT This is MoFo.

F

M $

A

M

J

J

A

S

O N $

D

J

F

M A $

17

Are Service Credits an Exclusive Remedy?  Service Credits can either be:  

a “bit off the price to reflect service not received” OR an attempt to compensate for all losses resulting from a Service Level failure

 Which is best? 

Most clients will prefer the best of both worlds i.e., the latter with an ability to claim actual loss as well

 No fixed rule of thumb

This is MoFo.

18

Response to Clients  The aim is to avoid Service Credits being punitive  Traditional best position = Service Credits as an exclusive remedy 

But beware that that might cause Customer to set Service Credits at a high level

 Reasons why it may be better to accept Service Credits as nonexclusive (as long as they are set low as a result)    

For all but “single event” Service Levels, damages and financial loss may be hard to prove Tendency for customers to be slow to claim or poor at keeping adequate records Probability of an intervening event Lack of direct linkage for periodic event Service Levels

 Work through probability of loss issues before choosing best approach This is MoFo.

19

Earnbacks What are they? Opportunity for providers to earn back previously paid Service Level Credits

Typical Customer View 1. undermines importance of hitting Service Levels each month 2. encourages over-performance 3. allows earn-back on services that the business doesn’t really care about 4. de-emphasizes importance of consistent performance

This is MoFo.

Provider View 1. Incentivizes recovery plans and remediation 2. Recognizes long-term nature of outsourcing relationship 3. More applicable to periodic Service Levels than one-off events of service failure 4. OK if targeted at subset of things that the business cares about

20

Earnbacks Do customers ever agree to earnbacks? Yes, but must be designed not to undermine Service Level effectiveness

What are appropriate parameters for earning Service Credits back?  No earnback for failure to meet the Minimum Performance  Earnback applies only on a Service Level for Service Level basis (i.e., cannot earn back a credit paid for one Service Level with good performance in another Service Level)  Earnback only applies if meet or exceed same Service Level for 12 consecutive months following the Service Level Failure  Can only earnback a portion of the Service Level Credit (e.g., 50%)

This is MoFo.

21

Operating Level Agreements (OLAs) Background Once Upon a Time …….. Customers outsourced Services using a Big Bang Approach But then …….. Customers wised up and realized that no single provider is the best at everything So …….. Customers began to choose multiple providers, picking the provider for a function that was best at performing that task But …….. This caused problems because all of the different providers had to play in the same sandbox Which meant …….. Somebody had to create rules for everybody to play nice

This is MoFo.

OLAs 22

OLAs – The Crux of the Problem Under Big Bang, provider was the integrator of the Service and Customer could look solely to the Provider for Service performance and failures In a multi-sourced environment, Customer IS the integrator and must hold each Provider accountable for its performance and failures

P1

P2

P3

Customer

P6

But when functions are related, there is overlap and multiple interrelations And where there is overlap, there is the possibility for confusion, which leads to a lack of accountability

P5

P4

OLAs are designed to create accountability by defining responsibilities among and between providers to reduce the finger pointing and claims of excused performance

This is MoFo.

23

OLAs – What are They? What is it? An OLA is an agreement between and among the customer and all of its providers whose functions are interrelated

What is its purpose? To create contractual privity between otherwise distinct providers, causing them to have shared responsibility for service delivery and creating more seamless overall service quality for the customer

Other Benefits: 

Clearly defines roles and responsibilities

  

This is MoFo.



Creates enforcement mechanism that keeps providers honest

Establishes a process for resolving multi-party disputes arising from interrelationships



Imposes “fix it first and argue later” culture

Creates a change management process for multiple providers



Imposes common IT and other standards

Avenue for sharing of necessary data and confidential information

24

OLAs – How Do They Work? First, a commitment in each of the base agreements between Customer and Provider that Provider will enter into an OLA with other related providers. At a minimum, that commitment must contain: (1) An outline for the structure of the OLA (2) An agreement on the types of confidential information that will be required to be shared among the various providers (3) Timelines for completing the OLA (4) Commitment to act and negotiate the OLA in good faith of which a failure is considered a material breach subject to termination of the base agreement Second, once the base agreements have been completed, the following steps to create a multi-party agreement: (1) Share Statements of Work (SOWs) under each base agreement with the other related parties (Note: may need to share Service Levels — but that is more difficult to negotiate) (2) After each entity has reviewed the other SOWs, engage in negotiation to agree upon a Dependencies schedule (i.e., where in order for one party to meet its Service or Service Level obligations, it is reliant on another party’s performance of its responsibilities) (3) Create excuse and penalty structure when Dependencies are missed This is MoFo.

25

OLAs – Content and Issues What is generally contained in the final OLA: (1) Agreement on sharing of Confidential Information and Intellectual Property (2) A Cross-Dependency Table (3) Agreement regarding Excused Performance

(5) Governance and Change Management Provisions (6) Agreement regarding cooperation business continuity during Disasters, Force Majeure Events (7) Dispute Resolution Mechanism

(4) Enforcement Mechanism What are some of the obstacles in getting to a final OLA?

     

Timing – negotiating with the applicable in parallel is optimal, but rarely happens Getting over the initial hurdle of sharing of confidential information Talking the same language Agreeing upon what is a Dependency Governing the OLA process Deal fatigue

This is MoFo.

26

OLAs – Enforcement for Failures Enforcement is a difficult proposition because one provider never wants to be liable for the failure of other providers. One provider is unlikely to pay damages to another provider. 

If Provider A fails to perform a task that causes Provider B to fail to meet a Service Level, what is the result without an enforcement regime?

1. Customer wants to hold somebody accountable



2. Provider B doesn’t want to be that somebody and wants Service Credit relief.

3. Provider A is trying to lay low – maybe its failure results in a Service Level Failure under its base agreement, maybe it doesn’t. If it doesn’t and Provider B gets relief, nobody is held accountable.

How do we resolve the problem?



Structure the penalty such that it is not paid to the other provider or based on penalty structures in the other party’s agreement



One solution is to make it a part of the Service Level regime under the base agreement.

Create a “Dependency” SLA

This is MoFo.

A failure of a Dependency SLA will occur when the provider fails to perform a function upon which another provider is dependent, as documented in the Dependencies document

The base agreement can define a Dependency failure such that there are Expected Targets and/or Minimum Levels – or simply, every time a Dependency Service Level is triggered a Service Level Credit occurs

27

Lunch & Learn

Alistair Maughan Partner, Technology Transactions T: (+44) 20 7920 4066 E: [email protected]

This is MoFo.

28

Suggest Documents