What Is Connection Latency and How Does It Affect Your Business?
When a business owner complains that the internet is slow, they usually mean one of two different things. Either the file took too long to download, or the video call kept cutting out, or the voice on the phone sounded like they were calling from underwater. These are different problems with different causes, and fixing one will not necessarily fix the other.
Speed and latency are both real metrics that affect how your business operates. Most people understand speed in rough terms. Latency is the one that gets overlooked — until your VoIP calls start sounding terrible or your cloud accounting software starts timing out in ways that your download speed test cannot explain.
Latency vs Speed — Why They Are Different and Both Matter
Speed (bandwidth) measures how much data can move through your connection per second. It determines how fast a large file downloads, how quickly a page full of images loads on first visit, and how many simultaneous users your connection can support without degrading.
Latency measures how long it takes for a small piece of data to make a round trip from your device to a server and back. It is usually measured in milliseconds (ms). The number you see in a ping test is latency — not speed.
Think of it this way. Speed is how wide the road is. Latency is how long the road is. You can have a very wide road (high bandwidth) that is also very long (high latency). On that road, large trucks (file downloads) eventually get through fine. But a conversation happening across that road — where both sides need to respond quickly to each other — suffers because every exchange takes too long to travel the distance.
Real-time applications — voice calls, video conferencing, remote desktop sessions — are almost entirely latency-sensitive, not bandwidth-sensitive. A VoIP call uses very little bandwidth. It needs very low latency and low jitter (variation in latency). A high-speed connection with poor latency will make your calls sound broken even while it downloads files quickly.
What Causes High Latency in Jamaica
Jamaica's geography creates a structural latency reality that every business on the island deals with. Most internet traffic that leaves a Jamaican network has to physically travel to a point of presence in Miami before it reaches the global internet. From Kingston to Miami, that cable runs approximately 1,100 kilometres under the Caribbean Sea.
The speed of light in fiber is roughly 200,000 kilometres per second, so the physical minimum for a round trip to Miami and back is around 11 milliseconds. In practice, network equipment, routing hops, and processing time add to this. A realistic baseline from Kingston to a Miami exchange point is 20–35ms. From Montego Bay, routing through Kingston first, you add a few more milliseconds.
When you are connecting to a server in the US — Microsoft 365 servers, Zoom, QuickBooks Online — you are adding the latency from that US location to the Miami base. A server in New York adds 10–15ms. A server in Dallas adds 20ms. European servers add 100ms or more.
CGNAT (Carrier-Grade NAT) compounds latency in a different way. Many Flow and Digicel business connections — particularly those on cable or mobile broadband — sit behind CGNAT, where multiple customers share a single public IP address. CGNAT adds a layer of address translation that introduces additional latency and can cause problems with specific protocols, including some VoIP configurations and site-to-site VPN connections.
If your ISP-supplied router shows a 100.x.x.x address or a 10.x.x.x address on the WAN port (not your LAN), you are likely behind CGNAT. This is not always solvable without upgrading to a dedicated IP plan, but knowing it explains certain connectivity behaviours that otherwise look like random problems.
How Latency Affects VoIP, Video Calls, and Cloud Applications
VoIP and Business Phone Systems
Voice over IP is the most sensitive common business application for latency. The ITU G.114 standard recommends one-way delay under 150ms for acceptable voice quality. Round-trip latency (what a ping test measures) should therefore be under 300ms to stay in the acceptable range.
Above 150ms one-way delay, callers start to notice slight lag. Above 250ms, the conversation becomes awkward — both parties start talking over each other. Above 400ms, the call sounds like a satellite connection from the 1990s.
Jitter (variation in latency) is often worse for VoIP than the absolute latency number. A connection with 80ms average latency but spikes to 200ms will produce choppy audio. VoIP endpoints use a jitter buffer to smooth this out, but a buffer large enough to absorb high jitter adds its own delay.
If your business phone system is cloud-hosted — which is the standard for modern PBX deployments — your call quality depends entirely on the path between your office and your VoIP provider's servers. A poor office internet connection, CGNAT, or a congested link to Miami will all show up as call quality problems.
Video Conferencing
Zoom, Teams, and Google Meet are more tolerant of higher latency than VoIP because they use more aggressive buffering. But above 150ms round-trip latency, video quality degrades, audio starts to desync from video, and participants notice dropped frames.
Packet loss is particularly destructive to video. Even 1–2% packet loss causes visible quality drops because the codecs have to skip or interpolate missing data. High latency and packet loss together make video calls effectively unusable.
Cloud Applications
Applications like QuickBooks Online, Sage, or any SaaS accounting or ERP platform make many small, rapid API calls as you navigate between screens. Each call has to complete before the next screen loads. With 30ms latency, navigating between screens feels instant. With 200ms latency, every click produces a perceptible wait. With 400ms latency, the application feels broken even though technically every request succeeds.
This is why a staff member with a 25 Mbps connection can feel slower than one with a 10 Mbps connection — if the 25 Mbps line has higher latency, cloud application responsiveness suffers despite the higher raw speed.
Acceptable Latency Thresholds for Business Use Cases
| Application | Acceptable round-trip latency | Maximum before noticeable degradation |
|---|---|---|
| VoIP phone calls | Under 150ms | 300ms |
| Video conferencing | Under 150ms | 200ms |
| Remote desktop (RDP) | Under 100ms | 150ms |
| Cloud applications (SaaS) | Under 100ms | 250ms |
| File transfers / backups | Not sensitive | Not applicable |
| Not sensitive | Not applicable |
For a Kingston business on a dedicated fibre connection to Flow or Digicel, latency to Miami-based servers should sit in the 25–50ms range. Latency to US East Coast servers should be in the 50–80ms range. If your measurements are significantly higher than these numbers, there is either a routing problem, congestion on your link, or a degraded line that warrants a support call to your ISP.
How to Measure Your Latency and Speed Right Now
You do not need any technical tools installed to get a useful picture of your connection.
Test your latency to the Jamaica Cloudflare edge. Cloudflare operates servers in Kingston, so this gives you a baseline that removes the Miami routing factor. Test your latency to the Kingston Cloudflare edge at CheckMiIP.com. A result under 20ms to the Kingston edge indicates a healthy local connection. High latency to a Kingston server suggests a problem between you and your ISP, not the broader internet path.
Run a download speed test. Run a download speed test at CheckMiIP.com to confirm your connection is delivering the bandwidth you are paying for. Compare the result to your plan speed. Significant underperformance (less than 70% of plan speed) is worth escalating to your ISP.
Check your full connection details. Check your full connection details at CheckMiIP.com to see your IP address, whether you are behind CGNAT, and which network you are routing through. This information is useful when troubleshooting VoIP or VPN problems.
For ongoing monitoring, a tool like PingPlotter can run continuous latency tests in the background and graph the results over time. This is invaluable for demonstrating to an ISP that a problem is real and recurring — a screenshot of a clean 5-minute test does not make the case that an intermittent problem at 3pm every day is costing your business.
When you have your latency measurements, compare them against the thresholds above. If your round-trip latency to a Miami server consistently exceeds 100ms, or if you are seeing jitter above 30ms, those numbers should drive a conversation with your ISP or with your IT support team about remediation options.
Measure your connection latency and speed at CheckMiIP.com — free network diagnostic tools built for Jamaica.
If your business is experiencing call quality problems, slow cloud applications, or VPN connectivity issues and you want someone to actually diagnose what is happening rather than guess, get in touch with Rubix Systems Jamaica. We support businesses across Kingston, Montego Bay, and across the island — and we will tell you exactly what your connection is doing and what can be done about it.