SLA-Backed On-Site IT Support in South Africa — What's Reasonable to Expect

Why on-site support still matters in a cloud-first world

Most IT is remote now — and it should be. But a warm body on site still resolves things nobody can fix through a screen: dead switches, cabling, printer meltdowns, boardroom AV failing 5 minutes before a client walks in, a laptop that won't boot after last night's power outage. If your business loses money when the office isn't working, on-site coverage is not optional.

Realistic SA response benchmarks

• Gauteng metros (JHB, PTA, Ekurhuleni). 4-hour on-site for P1 is standard SME; 2-hour is premium tier.
• Cape Town, Durban metro. Same 4-hour standard for P1 in-metro; outer suburbs slide to next business day unless explicitly priced.
• Other metros (Bloemfontein, PE, East London, Polokwane). Same-day in-metro for P1 is realistic; 4-hour requires a resident partner engineer.
• Rural / outlying. Next-business-day is the honest answer. Anyone quoting 4-hour on a farm outside the metros is either flying or fibbing.
• After-hours P1. 6–8 hours to on-site is the SME standard; faster than that = expect a 30–50% loading on retainer.

How to structure the retainer

1. Bundle it. Per-user/month covering remote + on-site + monitoring + patching + backup + endpoint security is almost always cheaper than à la carte call-outs beyond ~15 users.
2. Cap the on-site hours or make them unlimited — never leave "reasonable use" undefined. Ambiguity is where the disputes live.
3. Define coverage hours explicitly. Business hours (07:00–17:00 Mon–Fri) vs. extended (07:00–20:00 + Saturday) vs. 24/7. Each tier has real cost implications.
4. Ring-fence project work. Migrations, new office fit-outs and major changes are quoted separately — never allow them to eat the support retainer.
5. Include a monthly report — tickets by priority, SLA compliance %, top recurring issues, recommendations. If your MSP can't produce it, they can't manage what they can't measure.

Priority levels done right

• P1 — Business-critical. Server/internet/email/site down. Response: 15 min. On-site (if needed): 4 business hours. 24/7 coverage should apply.
• P2 — Significant impact. One team or department affected; printing broken; VoIP degraded. Response: 1 business hour. On-site: same business day.
• P3 — Single user broken. Response: 4 business hours. On-site or resolution: next business day.
• P4 — Request/change. New starter, permission change, small project. Response: next business day. Delivery: agreed per request.

Red flags in a proposed SLA

• "Best effort" response times. That means no commitment. Walk away.
• Only response time committed, no arrival time — a phone acknowledgment is not a fix.
• No escalation matrix with names and phone numbers.
• No credits, no penalties, no consequences if the SLA is missed — pure risk on you.
• Vague hardware and 3rd-party exclusions ("issues caused by external factors") — insist on specifics.
• No coverage of common SA realities: load-shedding, Eskom voltage events, ISP outages — these must be explicitly in scope with a defined role.

How to measure your MSP monthly

1. SLA compliance % per priority, per site.
2. Ticket volume trend — a good MSP's numbers should trend down as they fix root causes, not up.
3. Top 5 recurring issues — evidence of proactive work, not just firefighting.
4. Time to first response vs. time to arrival vs. time to resolution — the three numbers tell three different stories.
5. CSAT (customer satisfaction) sampled monthly across your end users, not just the IT contact.

Frequently asked questions