Field notes, from the point of view of someone who has had to explain a penalty invoice to a finance head. Short version: most operations managers know their SLA number and have never read the clause that turns that number into money. That is backwards. The clause decides what a miss costs. The dashboard only tells you that you missed.
What a penalty clause usually contains
Contracts differ, so treat this as a typical shape rather than a rule. A penalty, often called a service credit, is a reduction in what the client pays when agreed service levels are missed. It normally has five parts:
- The metric and its definition (service level, abandon rate, accuracy, turnaround time).
- The measurement period, usually a month, sometimes a quarter.
- Tiers: the bigger the miss, the bigger the credit.
- An at-risk cap: the most the client can claim, often a percentage of the monthly fee.
- Exclusions: events that stop the clock, such as client system outages or volumes above the forecast the client committed to.
My opinion: the definition and the exclusions are where accounts are won and lost, not the percentages. A penalty of 5 per cent on a metric you can game is less dangerous than 2 per cent on a metric measured in a way you did not understand when you signed.
A worked month
Take an illustrative account: 100 billable FTEs at ₹45,000 each per month. The monthly fee is ₹45,00,000. The contract targets 80 per cent of calls answered within 20 seconds, with these tiers on the monthly fee:
| Monthly service level | Credit | In rupees |
|---|---|---|
| 80% or above | None | 0 |
| 75% to below 80% | 3% | ₹1,35,000 |
| 70% to below 75% | 5% | ₹2,25,000 |
| Below 70% | 8% | ₹3,60,000 |
The at-risk cap is 10 per cent of the monthly fee, which is ₹4,50,000, so a bad month cannot exceed that regardless of how many metrics fail. Suppose you land at 77 per cent. The credit is ₹1,35,000. Now look at the decision this creates.
With ten days left and service level trending to 77, you could add three agents to the schedule for those days at, say, ₹35,000 each in monthly cost, which is ₹1,05,000 on a full-month basis and less if prorated. That might save ₹1,35,000. Paying extra staff to avoid a penalty is sensible if it works. It is not sensible if the real problem is a forecast that is wrong every Monday, because then you pay the extra every month. Penalties should push you to fix causes, not to buy insurance.
The cliff problem
Look at the table again. At 79.9 per cent the account pays ₹1,35,000. At 80.0 it pays nothing. A tiny difference in outcome produces a large difference in money. That is a cliff, and cliffs create bad behaviour. People start trimming the measured number instead of the real service: short-abandon rules get abused, callbacks are offered to stop the clock, difficult calls are quietly transferred. When you see a metric hovering at exactly the threshold for twenty days and then jumping, ask what changed in how it is counted.
Better clauses use gradual scales, or measure over a longer period so that one bad day does not sink a month. If you are the one negotiating, ask for that. If you are not, at least know where the cliffs are.
Earn-back and bonuses
Some contracts include earn-back: a credit paid in month one is returned if performance beats target for the next two or three months. This is useful because it rewards recovery. Others pair penalties with a bonus for exceeding target, for example a small percentage uplift above 85 per cent. Do the arithmetic on the bonus before celebrating. If hitting 85 needs staffing that costs more than the uplift, you are being paid to lose money.
Also check whether penalties and bonuses are symmetrical. In many outsourcing deals the penalty applies from the first point of miss while the bonus only starts well above target. That asymmetry is a pricing decision, whether anyone admitted it or not.
How to protect the account
- Record client-caused events the day they happen. A system outage that took 40 minutes of agent time cannot be argued about a month later without a log.
- Track forecast against actual volume. If the contract excludes volume more than a stated percentage above forecast, you need the evidence in a form the client accepts.
- Send a mid-month warning when a metric is at risk. Clients dislike surprises more than misses.
- Agree the calculation in writing, including how abandoned calls under a few seconds are treated.
I would also say this: do not negotiate penalties alone with the client’s operations contact. Involve finance and whoever owns the contract. A credit that looks small in operations language is a line item in someone’s budget.
Penalty versus relationship
A credit is not the true cost of a miss. The client may enforce it politely, waive it as goodwill, or hold it over you at renewal. Some clients never claim and still remember. Treat the penalty as the visible part of a larger bill and spend your effort on why the miss happened. Where the clause says you must provide a root-cause report within a set number of days, do it fast and make it honest. A clear explanation with a fix is often what turns a credit into a trust.
Last point. If you think the target itself was set wrongly, say so before you sign, not after the first penalty. A target the team cannot reasonably hit at the agreed price is not a stretch goal. It is a slow loss written into the contract.
Figures and names in the examples are illustrative.
Leave a Reply