If you have ever run an internet speed test showing 100 Mbps and then wondered why your downloads only reach about 12.5 MB per second, you are not alone. This is one of the most common points of confusion in IT, and it comes down to the difference between bits and bytes. A bit (b) is the smallest unit of data in computing — a single 0 or 1. A byte (B) is a group of 8 bits. The distinction matters because the networking industry historically measures data transfer rates in bits per second, while operating systems and file managers display file sizes in bytes. When your internet service provider (ISP) advertises a "100 Mbps" connection, they mean 100 megabits per second. To convert this to megabytes per second, you divide by 8: 100 Mbps ÷ 8 = 12.5 MB/s. This is the theoretical maximum download speed you would see in your file manager or browser download dialog.

Internet service providers use megabits per second (Mbps) rather than megabytes per second (MB/s) for several reasons. First, it is the established industry standard — networking equipment, protocols, and specifications have used bits per second since the early days of telecommunications. Ethernet standards, for example, are defined as 10 Mbps, 100 Mbps, 1 Gbps, and 10 Gbps. Second, using bits produces larger numbers. "100 Mbps" sounds faster than "12.5 MB/s" even though they represent the same speed. This is not necessarily deceptive — it is simply the convention — but it does contribute to consumer confusion. Third, network data is transmitted bit by bit over the wire. Each electrical signal or light pulse represents a single bit, making bits per second the most natural unit for measuring the rate at which data moves across a network link.

Even after converting from Mbps to MB/s, your actual download speeds will typically be lower than the theoretical maximum. Several factors contribute to this gap: Protocol overhead accounts for roughly 5-10% of your bandwidth. TCP/IP headers, Ethernet framing, and other protocol data consume bandwidth that does not contribute to your file download. For example, on a 100 Mbps connection, protocol overhead might reduce your effective throughput to about 90-95 Mbps, or approximately 11.2-11.9 MB/s. Network congestion occurs when many users share the same network infrastructure. During peak hours, your ISP's network may be more congested, reducing your available bandwidth. This is particularly common with cable internet, where you share bandwidth with your neighbors. Server-side limitations also play a role. The server you are downloading from may have bandwidth limits, high load, or geographical distance that reduces your download speed regardless of your connection capacity. Wi-Fi overhead is another significant factor. If you are connected wirelessly, Wi-Fi protocols add additional overhead, and signal strength, interference, and distance from the router all affect your actual speed. A Wi-Fi connection typically achieves 50-70% of the speed you would get with a wired Ethernet connection.

Here is a practical reference table showing common ISP speed tiers and what they translate to in real-world download speeds: • 25 Mbps plan → 3.125 MB/s theoretical → approximately 2.5-2.8 MB/s actual • 50 Mbps plan → 6.25 MB/s theoretical → approximately 5.0-5.6 MB/s actual • 100 Mbps plan → 12.5 MB/s theoretical → approximately 10.0-11.2 MB/s actual • 200 Mbps plan → 25 MB/s theoretical → approximately 20.0-22.5 MB/s actual • 500 Mbps plan → 62.5 MB/s theoretical → approximately 50.0-56.0 MB/s actual • 1 Gbps plan → 125 MB/s theoretical → approximately 100-112 MB/s actual These "actual" figures assume a wired Ethernet connection with typical protocol overhead. Wi-Fi connections will generally be lower.

To accurately measure your internet speed, follow these steps. First, connect your computer directly to your router or modem using an Ethernet cable. This eliminates Wi-Fi as a variable. Second, close all other applications and browser tabs that might be using bandwidth. Third, run a speed test using a reputable service like Speedtest.net, Fast.com (by Netflix), or Google's built-in speed test. The speed test will report your download and upload speeds in Mbps. To convert to the MB/s you will see in actual downloads, divide by 8. If the result is significantly lower than your plan's advertised speed (less than 80% of the theoretical maximum on a wired connection), contact your ISP to troubleshoot. For IT professionals managing networks, it is also useful to run iperf3 tests between two points on your network to measure actual throughput independent of internet factors. This helps isolate whether speed issues are internal (LAN) or external (WAN).

As gigabit internet becomes more common, the same confusion scales up. A 1 Gbps (gigabit per second) connection provides a theoretical maximum of 125 MB/s (megabytes per second), not 1 GB/s (gigabyte per second). This is a factor-of-8 difference that can lead to unrealistic expectations. Furthermore, most consumer devices cannot fully saturate a gigabit connection. Older hard drives max out at about 100-150 MB/s for sequential writes, meaning your storage device — not your internet connection — becomes the bottleneck. Solid-state drives (SSDs) can handle gigabit speeds easily, with NVMe SSDs capable of several GB/s. For IT professionals planning network upgrades, it is important to consider the entire data path: the internet connection speed, the router and switch capabilities, the Ethernet cable category (Cat5e for gigabit, Cat6 for 10 gigabit), and the storage device speeds at both endpoints.