SLA penalty clauses: what clients won't tell you, what the law won't let you do.

"R500 deducted for a late guard." It looks small. Across a 12-guard contract over 24 months, it is R20,000 to R80,000 of margin you eat. Section 34 of the BCEA forbids you from deducting that from the guard — so the business absorbs it. Here is what to negotiate before you sign, and what makes a penalty clause defensible instead of fatal.

The penalty that looks like nothing

Most service-level agreements bury a deductions table somewhere after the scope of work. A late shift start: R500. A guard who books off without a same-grade replacement: R1,000. An incomplete occurrence book: R250. Each line looks trivial against a contract worth hundreds of thousands of rand a year. That is exactly why operators sign it without modelling it.

Now put a frequency on it. A 12-guard contract is roughly 360 guard-shifts a month. If even 1% of those shifts trigger a R500 penalty — one late start in a hundred, which is optimistic for any real roster — that is R1,800 a month. Over a 24-month term:

That is R20,000 to R80,000+ over the life of a single mid-size contract — and it comes straight off your margin, which on a 24/7 post is roughly R3,000 per post per month to begin with. The penalty table can quietly consume an entire post's profit before you have accounted for a single fuel hike. Price the post honestly first with the pricing calculator, then look at the penalty table and ask whether any margin survives it.

Section 34: you eat it, not the guard

The instinct is obvious: if the guard was late, deduct the R500 from the guard. You cannot. Section 34 of the Basic Conditions of Employment Act tightly limits deductions from an employee's remuneration. You may not deduct for loss or damage unless the loss happened through the employee's fault, you followed a fair procedure and gave them a chance to respond, the amount does not exceed the actual loss, and total deductions stay within one-quarter of remuneration — and even then it requires written agreement.

A client's SLA penalty is not the guard's "loss or damage" in that sense — it is a commercial penalty between you and the client. It does not pass through to the worker. So the R500 hits your profit-and-loss, every time, with no recovery mechanism against the person who caused it. This is the single most misunderstood line in security contracting: the penalty is a business cost, not a payroll adjustment. (This is operator guidance, not legal advice — have your specific clauses reviewed.)

The three clauses that quietly convert a win into a loss

1. Uncapped penalties

The most dangerous version has no monthly ceiling. In a bad month — a resignation wave, a taxi strike, load-shedding wrecking your access control — penalties can exceed the invoice itself. You can end up paying the client to guard their site. Any penalty regime without a cap is an open-ended liability you cannot price.

2. Scope creep dressed as "reasonable instructions"

Watch for language like "the contractor will perform such additional duties as the client may reasonably require at no additional cost." That single sentence turns a fixed-price guarding contract into an elastic one. Today it is "also lock the back gate." In six months it is access control, parcel handling and after-hours patrols you never priced. Scope must be defined and finite, with a written variation-and-cost mechanism for anything beyond it.

3. Termination asymmetry

Read the exit terms from both sides. A common trap: the client can terminate on 30 days' notice for convenience, but you are locked in for 24 months or owe an early-exit penalty. That means all the term risk — wage increases, fuel, penalties — sits with you, while the client keeps the option to leave the moment you become inconvenient. Notice periods and exit rights should be symmetrical, or priced for the imbalance.

What a defensible penalty clause looks like

Penalties are not the enemy. A fair, capped SLA is a sign of a serious client and protects good operators from being undercut by cowboys. Negotiate toward these five features before you sign:

• A hard cap — total monthly penalties limited to, say, 5–10% of the monthly invoice. This is the single most important term. Get the cap and most of the danger disappears.
• A cure period — a defined replacement window (for example, 30 to 60 minutes to deploy a same-grade reliever) before a penalty triggers. Penalise unresolved failures, not the normal friction of running a roster.
• Material, measurable breaches only — penalties tied to things that actually matter and can be objectively logged, not subjective "presentation" or paperwork trivia.
• Mutual obligations — the client must provide site access, a usable guardroom, working ablutions and on-time payment. If their failures cause your breach, no penalty applies.
• Excusable events — load-shedding, civil unrest, strikes and other force-majeure conditions are carved out. You cannot deploy through a road blockade, and you should not pay for being unable to.

Price the risk you can't negotiate out

Sometimes the client will not move on the cap. Then you have a clear, unemotional decision. Add an SLA-risk line to your bid — the calculator gives you the compliant floor, but it deliberately excludes penalty exposure, so model it on top. Estimate a realistic monthly penalty at your true incident rate, add it to your cost base, and re-check your margin.

If the post still clears your 8% floor with the penalty risk priced in, bid it. If pricing the risk makes you uncompetitive and the penalties are uncapped, that is not a contract — it is a liability waiting for a bad month. Walk, and put the capacity behind a client who signs a fair SLA.