Sales Loves MarketingA service level agreement (SLA) is an agreement that sets measurable targets for a service and states what happens if the provider misses them. Most SLAs sit inside a contract between a vendor and a customer, but teams inside one company can use them too.
What is an SLA used for?
An SLA turns a vague promise ("reliable service", "fast support") into numbers both sides can check. It answers three questions: what is being delivered, how it will be measured, and what the customer gets if the target is missed.
Software and cloud vendors publish SLAs for uptime and support response times. Outsourcing contracts use them for things like ticket resolution. Large companies also write internal SLAs between departments, so one team knows what it can expect from another.
What goes into an SLA?
Most SLAs cover the same building blocks:
- The service. What is included, and what is excluded.
- Targets. The service levels promised, such as monthly uptime or first response time.
- Measurement. How each target is measured and reported, and over what period.
- Reporting issues. How the customer raises a problem and how quickly it is handled.
- Remedies. What happens if a target is missed, such as service credits, a right to terminate, or an escalation path.
Common metrics include uptime, response time, resolution time and mean time to repair.
SLA vs SLO vs SLI
Google's Site Reliability Engineering book separates three terms:
| Term | What it is | Example |
|---|---|---|
| SLI (indicator) | The measurement | Share of minutes the service is reachable |
| SLO (objective) | The target for that measurement | 99.9 percent per month |
| SLA (agreement) | The contract that holds SLOs and the consequences of missing them | Service credits if uptime drops below 99.9 percent |
Google's test is simple: if missing the target has no explicit consequence, you are looking at an SLO, not an SLA.
What does an uptime SLA look like?
Google Cloud's Compute Engine SLA is a public example. In most regions it promises at least 99.99 percent monthly uptime for instances spread across multiple zones and 99.9 percent for most single instances. If Google misses the target, the customer can claim a credit on future bills: 10 percent of the monthly charge for a small miss, 25 percent below 99 percent uptime, and 100 percent below 95 percent.
Each extra "nine" matters. In a 30-day month, 99.9 percent allows about 43 minutes of downtime and 99.99 percent about 4.3 minutes. Across a full year, 99.999 percent allows only about 5.3 minutes.
How do sales and marketing teams use SLAs?
In B2B companies, the same idea is often applied to the lead handoff. Marketing commits to a number of qualified leads per month, defined in a way sales has agreed to. Sales commits to contacting each lead within a set time and recording the outcome in the CRM. Both sides then track the agreed KPIs together.
The value is the shared definition. When marketing and sales agree in writing on what a qualified lead is and how fast it gets a call, arguments about lead quality turn into numbers both teams can see. If you use one, review it every quarter, because lead sources and sales capacity change.
Related terms
Frequently asked questions
What is the difference between an SLA, an SLO and an SLI?
An SLI is the measurement, such as the share of requests that succeed. An SLO is the target for that measurement, such as 99.9 percent per month. An SLA is the agreement that contains SLOs and spells out the consequences of missing them, such as service credits.
What does 99.9 percent uptime mean in practice?
It allows about 43 minutes of downtime in a 30-day month. Each extra nine cuts the allowance by a factor of ten, so 99.99 percent allows about 4.3 minutes a month and 99.999 percent about 5.3 minutes a year.
What is an SLA between sales and marketing?
It is an internal agreement in which marketing commits to delivering a number or quality of leads and sales commits to following up on them within an agreed time. It gives both teams the same definition of a qualified lead and a shared way to measure the handoff.