Service level explained — what 80/20 means and where it comes from
Service level is the single most important metric in inbound contact centre operations. It drives your staffing requirement, your cost per contact, and your customers' experience. Yet most operations run a target they inherited without knowing where it came from.
What is service level?
Service level = the percentage of inbound contacts answered within a target time threshold. Written as X% in Ys, for example 80% in 20 seconds.
If 800 of 1,000 calls in an hour were answered within 20 seconds, service level = 80%. The other 200 either waited longer, abandoned, or, in some definitions, are excluded from the calculation.
Why 80% in 20 seconds?
The 80% in 20 seconds target is often called an “industry standard”, but it has no single scientific basis. It became a convention through industry adoption. AT&T popularised it in the 1980s for their consumer services, and it spread as a default benchmark.
In practice, the right service level target depends on your operation, your customers, your competitive environment, and often your contract or regulatory obligations. Some contact centres with complex, high-value calls use 80% in 60 seconds. Emergency lines target 95% in 10 seconds. Internal helpdesks use 80% in 2 minutes.
The practical implication
Changing from 80% in 20s to 80% in 30s can reduce your staffing requirement by 5–15% depending on traffic intensity. Raising from 80% to 90% at the same threshold typically requires 10–20% more agents. The cost difference is real — use the Erlang C calculator to model the headcount and cost trade-off.
How service level is calculated
In planning (forward-looking), service level is calculated using the Erlang C formula. Given a traffic intensity (Erlangs = calls × AHT ÷ interval length) and agent count, Erlang C computes the probability that a call will have to wait, and from that, the probability of being answered within any time threshold:
Where N = agents, A = traffic intensity (Erlangs), t = answer-time threshold, C(N,A) = Erlang C probability.
In reporting (backward-looking), service level is measured directly from your phone system or ACD: how many calls landed and how many were answered within the threshold.
Service level vs. ASA — why the difference matters
Service level is a threshold measure — what proportion of contacts met a time target. ASA (Average Speed of Answer) is a mean — what was the average wait time.
A contact centre can have a good ASA but a poor service level. If 95% of calls are answered in 2 seconds but 5% wait 15 minutes, ASA might be 50 seconds, which looks reasonable. But service level (80% in 20s) is 95%, which is excellent. Or flip it: if 50% wait zero seconds and 50% wait 2 minutes, ASA = 60s, service level = 50% in 20s.
Use service level, not ASA, for staffing decisions
ASA disguises long-tail wait times. Service level is honest about the proportion of customers who had a bad experience. Staff to service level; use ASA as a secondary diagnostic.
Service level benchmarks by sector
Benchmarks are operational conventions, not regulatory requirements unless specified.
Service level — frequently asked questions
What is service level in a contact centre?
Service level is the percentage of inbound contacts answered within a target time threshold. It is expressed as X% in Y seconds. For example, '80% in 20 seconds' means 80% of calls are answered within 20 seconds. The remaining 20% may wait longer or, if callers abandon, not be answered at all.
Why is 80/20 the standard service level target?
The 80% in 20 seconds target is often cited as a 'standard', but it has no single authoritative origin. It became a convention through industry adoption rather than scientific study. The 80/20 split was popularised by AT&T in the 1980s and became an industry default. Many organisations now set targets based on their specific customer experience goals, contract requirements, or regulatory obligations rather than the generic benchmark.
What is the difference between service level and ASA?
Service level is a percentage — what proportion of contacts met a time threshold. ASA (Average Speed of Answer) is an average — what was the mean wait time across all contacts. They measure different things. Service level is generally the better KPI: a low ASA can hide a long tail of very long waits. If 95% of callers are answered in 5 seconds but 5% wait 10 minutes, ASA looks fine but service level reveals the problem.
What happens to staffing requirements when I raise my service level target?
Raising your service level target from 80% to 90% typically requires 5–20% more agents, depending on current occupancy and traffic intensity. The relationship is non-linear: going from 70% to 80% SL is cheaper than going from 90% to 95%. At very high targets (95%+), the marginal cost per percentage point rises sharply. Use the Erlang C calculator to see the exact agent count required for any target.
Related
Erlang C calculator
Agents required to hit your SL target
AHT impact calculator
How handle time affects your SL
Erlang C explained
How the staffing formula works
Occupancy calculator
The SL and occupancy trade-off
Service level in the glossary
Quick reference definition
ASA in the glossary
Average Speed of Answer explained
Model your service level target
See exactly how many agents you need to hit 80%, 85%, or 90% service level — and what it costs.
Open the free Erlang C calculator →