An SLA is the promise you make to customers, and an OLA is the internal commitment between teams that makes that promise achievable. They are complementary, not competing. Underpinning contracts with outside vendors must line up with both, because a service target is only as strong as the weakest handoff behind it.
TL;DR:
- Set customer targets from ticket history, define how each metric is calculated, and review results regularly before tightening commitments operations cannot yet meet.
- Name an owner for every internal task, set team handoff and acknowledgment windows, and document escalation paths for stalled tickets.
- When a vendor contract falls short of an SLA target, renegotiate its terms, name an escalation contact, or revise the customer commitment.
- Track service availability, customer response time, and restoration time in SLAs; measure assignment delays, queue time, and team handoffs in OLAs.
Table of Contents
- What is a service level agreement (SLA)?
- What is an operational level agreement (OLA)?
- Key differences: SLA vs OLA
- How SLAs, OLAs and underpinning contracts work together
- KPIs, metrics, and examples: measuring SLAs and OLAs
- Practical checklist and template elements for drafting SLAs and OLAs
- Author perspective: practical priorities for IT service managers
- How a managed IT partner can help implement SLAs and OLAs
- FAQ
- Sources
What is a service level agreement (SLA)?
A service level agreement is an agreement between an IT service provider and a customer that documents service level targets and responsibilities. The signatories are typically the provider’s service owner and a business-side representative of the customer. Because it faces outward, it is the document your customers point to when they ask what they were promised.
Most SLAs share a common set of elements:
- Service description: what is covered, and what is not.
- Service level targets: measurable commitments such as availability or response time.
- Responsibilities: what the provider does and what the customer must do.
- Reporting cadence: how often results are shared and in what format.
- Remedies or credits: what happens when a target is missed.
The SLA sets customer expectations in plain terms. It also gives both sides a shared yardstick for judging whether the service is usable, dependable, and pleasant to work with, rather than arguing from memory after an incident.
Your help desk may be excellent, but if the customer never agreed on what “excellent” means, every missed ticket becomes a debate. Writing the target down ends that debate before it starts.
What is an operational level agreement (OLA)?
An operational level agreement is an internal contract between departments. OLAs specify the internal activities and relationships required to meet the requirements of a related SLA. The signatories are internal teams, such as the service desk, the network team, and the security group.
A solid OLA usually includes:
- Internal tasks: the specific work each team owns, such as networking or security tasks.
- Handoff timings: how quickly a ticket must be accepted and moved along.
- Escalation paths: who gets called, and when, if work stalls.
- Responsibilities: a named owner for each step.
OLAs matter most where silos form. A service desk can promise a restore within an agreed window, but if the infrastructure team treats those tickets as low priority, the promise fails inside the building. An OLA turns that hidden dependency into an enforceable internal commitment, so each team knows what the others are counting on.
Key differences: SLA vs OLA
The simplest way to separate the two is to ask who the document is for and who answers when it is missed.
- Audience and purpose: an SLA faces the customer and sets expectations. An OLA faces your own teams and coordinates work.
- Scope and signatures: an SLA is a contract or formal agreement with a customer. An OLA is an internal agreement between departments.
- Enforcement and remedies: SLA misses often trigger customer credits or formal reviews. OLA misses go through internal governance, such as management review and process fixes.
- Examples: an incident response SLA promises a customer a first response within a set window. The matching OLA says the network team must accept an escalated ticket within its own set window.
Practitioner summaries describe SLAs as outward-facing and OLAs as inward-facing, and say both should be synchronized so service levels are actually met.
Think of it this way: the SLA is the headline, and the OLA is the plan behind it. If you only write the first, you have a promise without a method.
How SLAs, OLAs and underpinning contracts work together
An underpinning contract (UC) is an agreement between an IT service provider and a third-party vendor. According to Matrix42’s documentation, a UC defines the services and performance targets required from external suppliers to support the service promised in the SLA. Put simply, the SLA is the promise to the customer, the OLA is the internal plan to keep it, and the UC is the external support that helps fulfill it.
The typical flow runs in one direction:
- SLA requirement: the customer-facing target is agreed first.
- Internal OLAs: each team receives its share of the work and its timing.
- UC with vendors: outside suppliers commit to the pieces you cannot deliver alone.
When you rely on outside providers, your contract defines scope and service levels, and your internal teams focus on managing those obligations and interactions. Our vendor risk management guide covers the supplier side in more depth.
Sometimes a UC falls short of what the SLA needs. You have three honest options: use change control to renegotiate the vendor terms, set up an escalation path with a named vendor contact, or revisit the SLA target itself. Quietly absorbing the gap is a poor choice, because the customer may eventually feel it.

KPIs, metrics, and examples: measuring SLAs and OLAs
Good metrics are specific, countable, and tied to something the business cares about. Start with the customer-facing numbers for SLAs:
- Uptime or availability percentage: how much of the agreed time the service works.
- Mean time to restore (MTTR): the average time to bring a failed service back.
- Response time: how quickly a person acknowledges a request.
Then measure the internal steps for OLAs:
- Mean time to assign: how long a ticket waits before an owner is named.
- Handoff time: how long a ticket takes to move between teams.
- Queue time: how long work sits before anyone touches it.
- Internal acknowledgment time: how fast a receiving team confirms it has the work.
Choose targets you can actually hit. Service level management guidance says SLAs work best when they are realistic, aligned to customer needs, and reviewed regularly. Look at your own ticket history first, then set the target just beyond today’s performance. Our SLA metrics guide walks through sample commitments.
Pro Tip: Set targets in small steps, then review them regularly. Tightening a target after you have proven you can meet it builds trust. Promising an aggressive number on day one builds a backlog of apologies.
Practical checklist and template elements for drafting SLAs and OLAs
Use these steps as a starting point for your own documents. The first five cover SLAs and the rest cover OLAs.
- Define the service scope, including what is excluded.
- Set measurable targets with a clear way to calculate each one.
- Choose a reporting format and schedule.
- Agree on a review cadence and the remedies for missed targets.
- List what the customer must provide, such as access and contacts.
- Name an internal owner for every task in the OLA.
- Set handoff timings between teams and an escalation matrix.
- Record dependencies on vendors and the UCs that cover them.
- Add test and verification steps so you can confirm each handoff works.
For wording, keep it plain. A handoff clause might read: “The network team acknowledges escalated tickets within the agreed window and names an owner.” A responsibility clause might read: “The service desk owns customer communication until the ticket is closed.” Our list of questions to ask an IT provider is also useful when you need to check a partner’s side of the process.
Author perspective: practical priorities for IT service managers
Most SLA trouble starts when the document is written to catch someone out instead of to help the business. Tie every target to an outcome people can feel, such as staff getting back to work quickly. Keep agreements simple and measurable, and treat them as a working relationship, as ITIL guidance does. We hold ourselves to this: our help desk is U.S.-based, and we focus on responding promptly. That operational focus beats a thick contract nobody reads.
— Jeffrey
How a managed IT partner can help implement SLAs and OLAs
If you would rather not build this machinery yourself, we can help. Our managed IT services can help you align targets with your goals. When you evaluate any partner, look for a U.S.-based help desk and response times they can measure and report. Our provider scorecard lays out the criteria. Contact us to get started.

FAQ
What is OLA in ITIL?
An OLA, or operational level agreement, is an internal agreement between teams within the same organization. It spells out who does what, and how quickly, so the team can meet the targets promised in the customer-facing SLA.
What is the difference between SLA and OLA and underpinning contract?
An SLA is the promise to the customer, an OLA is the internal plan to keep that promise, and an underpinning contract is the agreement with an outside vendor that supports it. Each one covers a different party, and all three should agree on the same targets.
What are the three types of SLAs?
Commonly described SLA categories include customer-based, service-based, and multi-level (or tiered) agreements. A customer-based SLA covers one customer group, a service-based SLA covers one service for everyone, and a multi-level SLA layers both. Definitions vary slightly between organizations, so check how your own documents divide them.
How do you measure SLA and OLA performance?
Track customer-facing numbers such as availability, response time, and time to restore for SLAs. For OLAs, track internal numbers such as time to assign and handoff time. Review both on a regular schedule and adjust targets when your results show you can do better.
Sources
- Docs
- SLA, OLA and UC – Tahoe SP1 – Technical Documentation
- What Is an Operational-Level Agreement (OLA)? | Definition from TechTarget