How to Run an Internet Speed Test and Interpret Results to Improve Your South African Website

Picture this: you’re a small business owner in Johannesburg. Your fibre connection feels fine, your router is blinking happily, and yet your website loads painfully slowly on your phone. You refresh. Still slow. You ask a colleague to check it. Same result. So what’s actually wrong?

The honest answer is: you don’t know yet. “Fine” is not a measurement. Without running a proper internet speed test, you’re guessing at the problem, which means you’re also guessing at the solution. Speed testing is not just about confirming that you’re getting your subscribed Mbps. It’s about diagnosing exactly where the bottleneck lives: your ISP connection, your Wi-Fi, your router, or your hosting environment. Each one has a different fix. Running an internet speed test correctly, and knowing how to read the results, is the difference between a targeted fix and weeks of trial and error.

Tools like Speedtest by Ookla are free and take under a minute to run. The problem is that most people run them once, glance at the number, and stop there. By the end of this article, you’ll know how to choose the right tool for the right situation, run tests that produce results you can actually trust, read the numbers against meaningful benchmarks, and trace slowness to its true source. If that source turns out to be your hosting, there’s a direct path forward there too.

What an Internet Speed Test Actually Measures

Before you run a single test, it helps to understand what you’re measuring. Most people fixate on one number and ignore the rest, which leads to misdiagnosis.

Download Speed, Upload Speed and Why Both Matter for Your Business

Download speed measures how fast data travels from the internet to your device, expressed in megabits per second (Mbps). Upload speed measures the reverse: how fast data leaves your device and travels outward. A higher number in either direction means more data moves in that time.

For most casual browsing, download speed is the dominant factor. But if your business uploads product photos to an e-commerce platform, sends large files via email, or hosts video calls, your upload speed matters just as much. A connection advertising 100 Mbps download but only 5 Mbps upload will bottleneck you the moment you start sending rather than receiving.

Ping, Latency and Jitter: The Numbers People Ignore

Ping measures the round-trip time in milliseconds (ms) between your device and a test server. Latency describes the same concept more broadly: it’s the delay between a request being sent and a response arriving. Jitter measures how inconsistent that delay is from one moment to the next.

All three numbers are critical for any real-time use: video calls, live chat, and online gaming all suffer when ping is high or jitter is erratic, even when download speeds look healthy.

Here’s a detail that’s especially relevant in South Africa: a visitor hitting your website on a server in Amsterdam will always experience higher latency than a visitor hitting a server in Johannesburg, even with identical Mbps. Physical distance adds unavoidable milliseconds, and no amount of bandwidth compensates for that delay. This is why server location is not a minor technical detail.

TTFB: Bridging Your Connection Speed to Your Website’s Performance

Time to First Byte (TTFB) is the gap between a visitor’s browser sending a request to your website and the first byte of data arriving back.

A standard internet speed checker doesn’t measure TTFB directly, but understanding your ping and latency numbers helps explain it. If your speed test to a local Johannesburg server shows low ping but your website still feels sluggish, TTFB is the next diagnostic step.

For a South African site hosted locally, a TTFB notably above 300–500 ms is a concern and warrants investigation into server processing and hosting infrastructure. Sustained TTFB above 600–800 ms is especially likely to indicate server-side or hosting-network issues rather than anything to do with the visitor’s connection.

If your internet connection is performing well but your website is still slow, our guide on why your website is slow looks at some of the other common causes of poor website performance.

Choosing the Right Internet Speed Test Tool for South Africa

Not all tools measure the same things, and not all of them are equally useful for diagnosing local performance. Using the wrong one can give you results that look great on screen but tell you nothing useful about your actual situation.

The Free Tools Worth Bookmarking

Five tools are worth keeping in your diagnostic toolkit, each suited to a slightly different scenario:

• Speedtest by Ookla – the most widely recognised option, with thousands of servers across 190-plus countries. It measures download, upload, ping, and jitter, making it a reliable general-purpose ISP benchmark.
• Fast.com – simpler and ad-free, running on Netflix infrastructure. Particularly useful for checking streaming-style download performance.
• Cloudflare Speed Test – goes deeper on network quality by reporting latency, jitter, and packet loss. The best choice when video call quality or gaming lag is your concern.
• TestMy.net – uses HTML5 and keeps historical graphs. Because it doesn’t share infrastructure with major ISPs, it’s worth running as an independent cross-check.
• M-Lab – open-source and research-grade, useful when you want verifiable measurements for a formal complaint.

Use Ookla for a familiar baseline, Cloudflare when call quality is the issue, and TestMy.net or M-Lab when you want a result that no ISP can be accused of influencing.

Why Your Test Server Location Changes Everything

This is the most commonly misunderstood variable in South African speed testing. If you select a server in London or Amsterdam, you’re introducing international latency into your result, and that latency has nothing to do with the quality of your local fibre connection.

You’re essentially measuring an international route, not your access line. Always select a Johannesburg or Cape Town server to get a result that reflects your actual local connection performance. This step is especially important when you’re trying to understand the experience your South African website visitors are having.

For a deeper explanation of how server location affects website performance, read our guide on why South African hosting is faster for local visitors.

How to Run an Internet Speed Test You Can Actually Trust

Most browser- or app-based tests complete in roughly 30 to 90 seconds. Running one that produces reliable, repeatable results takes a few deliberate choices. Skip these steps and your results will be noise, not data.

Testing on a Wired Ethernet Connection

A wired Ethernet connection is the most reliable baseline for your fixed broadband service. Plug a laptop directly into your router using a Cat5e or Cat6 cable, disable Wi-Fi on that device, and pause everything that uses bandwidth: cloud backups, streaming services, VPN connections, and software updates.

Run at least three consecutive tests and record the median result, not the best one. This wired number gives you a useful baseline. Your Wi-Fi and wireless devices can never reliably reproduce the same conditions, which is why establishing a wired result first matters so much.

Testing on Wi-Fi and Mobile Data

For Wi-Fi, use the same device and the same test tool. Run the first test close to the router, then move to your usual working location and run it again. The gap between those two results tells you whether the problem is your ISP or your wireless coverage.

If the close-to-router result matches your Ethernet baseline but drops sharply in another room, you have a Wi-Fi coverage problem, not necessarily an ISP problem. These two issues have completely different fixes.

For mobile data, turn off Wi-Fi completely, confirm the phone displays a mobile signal type (3G, 4G/LTE, or 5G), disable VPNs and background downloads, and stay in one fixed location. Run three tests and record the median.

A 5G indicator on your phone does not guarantee strong performance; congestion, indoor obstruction, and tower distance all affect the actual result.

Mistakes That Make Results Unreliable

The most common errors that skew a broadband speed test are straightforward to avoid: running only one test and treating it as definitive, testing over a VPN, comparing tests from different servers as though they’re equivalent, and testing during peak congestion hours without knowing it.

Another frequent mistake is running a speed test while other devices on the network are streaming or uploading in the background. Any one of these will produce a result that misrepresents your actual connection, which leads to the wrong conclusion and the wrong fix.

Reading Your Results: What the Numbers Mean for a South African Business

A number without context is meaningless. The figures below cover the most common business and household scenarios, giving you benchmarks that make your results actionable rather than abstract.

Download, Upload and Ping Thresholds for Common Tasks

For general browsing and email, 5 to 10 Mbps download with ping under 100 ms is workable. HD video calls need around 10 Mbps download and 3 to 5 Mbps upload, with ping under 50 ms. HD streaming requires roughly 10 Mbps per simultaneous stream; 4K streaming jumps to 25 to 50 Mbps per stream.

For competitive online gaming, latency is the dominant factor: under 20 ms is excellent, under 50 ms is good, and above 100 ms is noticeably laggy regardless of how high your download speed is.

For a household running HD streaming, video calls, and browsing simultaneously, a 100 Mbps fibre plan is a practical minimum. Multiple 4K streams or several active users push that requirement toward 200 to 300 Mbps.

Spotting Congestion Versus Consistent Underperformance

If your speeds are reasonable at 07:00 but collapse between 18:00 and 22:00, that’s peak-time congestion, typically in the ISP’s network or the local fibre infrastructure.

This is not necessarily a device problem or a Wi-Fi problem. If speeds are consistently poor at all hours on a wired Ethernet connection, the cause is more likely a last-mile line issue, a faulty ONT, or a mismatch between your plan and what your line can actually deliver.

Test at three different times across two days and record each result to see the pattern clearly before drawing any conclusions.

When Your Internet Is Fine but Your Website Is Still Slow

This is where internet speed test results connect directly to your website’s performance, and where the diagnosis gets specific.

What Slow TTFB Tells You About Your Hosting

If your broadband speed test shows strong download speeds and low ping to a Johannesburg server, but your website still loads slowly for South African visitors, the bottleneck may be the hosting environment.

A TTFB notably above 300–500 ms for a locally hosted South African site is worth investigating. Sustained TTFB above 600–800 ms can suggest that the server is taking too long to generate and send a response, although other factors should also be checked.

For more information on website performance and hosting problems, read our guide on common hosting mistakes South African businesses make.

Images, Caching and Content Delivery

Once you’ve ruled out hosting response time, two other culprits are responsible for much of the remaining slowness. Uncompressed images are a common offender: a 5 MB hero image adds unnecessary data to every first-visit page load, regardless of connection speed.

Compress images, resize them to their displayed dimensions, and use modern formats like WebP.

Poor caching is another issue. Without proper cache rules, your server may need to rebuild pages from scratch on every request instead of serving efficiently cached versions. These problems compound each other, and they can compound the effect of slow hosting.

A clean speed test result helps you confirm the visitor’s internet isn’t the problem, which points the diagnosis directly at the website’s technical setup.

Why Local Hosting Changes the Equation for South African Visitors

When a website is hosted on a server in South Africa, the physical distance between that server and a local visitor is dramatically shorter than routing to Europe or the US. That distance reduction can reduce latency and improve response times without any changes to the site itself.

Hostigo’s guide on South African hosting and local visitor performance explains how server proximity affects latency and website performance.

For South African businesses, choosing infrastructure that is geographically aligned with your target audience can therefore be an important part of a broader website-performance strategy.

After any hosting change, run your internet speed checker again against a local server and measure your TTFB. The numbers will help show whether the move has made a difference.

How to Build a Complaint Case Against Your ISP

If your wired Ethernet tests consistently fall well below your subscribed plan speed, you may have grounds to raise a formal complaint. A single screenshot won’t provide much context. A well-documented evidence pack will be far more useful.

What to Record and How

For each test, record the exact date and time, download and upload speeds, ping, jitter, and packet loss, the test platform used, the specific South African server selected, the device and connection type (Ethernet only, not Wi-Fi), and whether other devices were active on the network.

Collect at least five wired tests across three days, including at least two during peak evening hours. A timestamped CSV export or spreadsheet log is more useful than a single screenshot because it shows whether the problem is consistent.

Also retain the contract or service agreement that specifies the advertised speed, any fair-use terms, and your billing records.

Escalating to ICASA When the ISP Doesn’t Act

South African consumers should complain to their ISP first and follow the applicable dispute-resolution process before escalating the matter. If the ISP fails to resolve the issue or responds unsatisfactorily, you can review the Independent Communications Authority of South Africa (ICASA) consumer complaints procedure.

State the advertised speed versus the measured performance, attach your representative test logs, include relevant ISP correspondence and reference numbers, and provide a clear timeline of when the problem began, what tests were run, and your interactions with the provider.

What to Do After You’ve Run the Tests

Running an internet speed test correctly takes a few extra minutes, but it gives you something genuinely valuable: clarity. You stop guessing and start diagnosing.

If your connection tests well on Ethernet against a Johannesburg server but your website is still slow for South African visitors, you now know that your internet connection may not be the main problem. Your hosting environment, website configuration, images, caching and other technical factors should then be investigated.

If hosting appears to be the issue, our guide to hosting for South African small businesses explains some of the factors to consider when choosing hosting, including speed, uptime, SSL, backups and server location.

Start by measuring your current site’s TTFB from a South African location using an internet speed checker or a dedicated TTFB tool, then compare it against a locally hosted alternative. If your hosting infrastructure isn’t built with South African visitors in mind, it’s worth addressing directly.

Retest after making changes. The numbers will help tell the story clearly.