Mbps vs MB/s: Why Downloads Are Slower Than Advertised

Somewhere between the plan on the sign and the progress bar on the screen, most people lose about a factor of eight, and almost nobody explains why. You pay for 100 Mbps and the download client shows 11 MB/s, and the gap looks like a betrayal. It is actually a unit conversion, followed by a series of small, physical costs of doing business on a real network. This guide separates the two quantities hiding inside that one-letter difference - megabits and megabytes - walks the arithmetic that connects them, and then accounts honestly for every remaining slowdown: protocol overhead, Wi-Fi losses, shared connections and servers that set the pace. By the end, the number in your download client will read not as a disappointment but as the expected answer to a solvable question.

CHAPTER 01Two Units Separated by One Letter

The lowercase b and the uppercase B carry the entire confusion. Mbps - megabits per second - is the unit internet providers use, because telecommunication has always counted bits, the individual 1s and 0s of transmission. MB/s - megabytes per second - is the unit file systems use, because storage has always been grouped into bytes of 8 bits for practical convenience. Both units are legitimate; they simply describe the same traffic at two different granularities.

The bridge between them is fixed and universal: 8 bits make 1 byte. That single fact converts any speed from one unit to the other. Divide bits per second by 8 to get bytes per second, or multiply bytes per second by 8 to get bits. A 100 Mbps line therefore cannot exceed 12.5 MB/s, because 100 divided by 8 is 12.5. Nothing about your router, provider or computer can change that ratio - it is arithmetic, not performance.

Why do two conventions survive side by side? History and marketing both claim credit. Bits were the native currency of early telecommunications, and a bigger number reads better on an advertisement: 100 Mbps and 12.5 MB/s describe the identical capacity, and the first one is the number that sells the plan. Knowing the conversion turns the advertising choice from a trap into a footnote, which is the real reward for learning the difference.

CHAPTER 02The Eight-Bit Bridge in Action

Put the conversion to work on a real example. A 700 MB file contains 5,600 megabits of data. On a 100 Mbps connection, moving 5,600 megabits at 100 megabits per second takes 56 seconds - the ideal, theoretical time. Notice what the byte-based view said: 700 MB at 12.5 MB/s also gives 56 seconds. The two routes agree, because they are the same route with the 8 already folded in at different points.

The free Bandwidth Calculator automates exactly this chain - size to bits, divide by speed, report the time - and it also shows the throughput line, where a 100 Mbps selection reads out as 12.50 MB/s. Seeing both numbers on one result screen cements the relationship faster than any explanation: the advertised figure and the client-side figure are one quantity wearing two units, and the calculator is the translator.

A third useful figure falls out of the same math: hourly capacity. At 100 Mbps, an hour moves 3,600 seconds times 100 megabits, or 360,000 megabits - about 45 GB. Per-hour thinking is often the most intuitive of all, because it turns an abstract speed into a volume you can picture: a streaming-heavy evening, a game update, a backup. Every speed converts this way, and the differences between plans become differences in how much data an hour of connection is worth.

CHAPTER 03Advertised Speed Is a Ceiling, Not a Promise

Even after the correct unit conversion, downloads rarely hit the theoretical maximum. The honest expectation for real transfers is somewhere between 60 and 90 percent of the advertised speed. The plan's headline number describes the line's raw capacity in ideal conditions; a working transfer spends some of that capacity on bookkeeping before any of your file arrives.

The bookkeeping has names. Every packet of data travels wrapped in headers - addressing, sequencing, error-checking - and those headers are overhead, real bits moving but not carrying file content. Encryption adds another layer. Acknowledgment traffic flows the other way, consuming upstream capacity while the download runs. Individually each cost is small; stacked together they commonly consume a tenth or more of the pipe, which is why practitioners planning transfers often assume a typical real-world figure at about 75 percent efficiency.

That planning figure converts directly into time. The 700 MB file with an ideal 56-second download lands nearer 75 seconds at 75 percent efficiency. A useful rule of thumb: expect real times roughly 1.1 to 1.7 times the ideal. The low end belongs to wired connections talking to fast nearby servers; the high end belongs to Wi-Fi in a crowded building or a server that limits each connection. When a transfer exceeds even the pessimistic end consistently, the cause is specific and findable, not mysterious.

CHAPTER 04Where the Missing Speed Actually Goes

After overhead, the next suspect is the wireless link. Wi-Fi shares open airspace with neighboring networks, Bluetooth devices, microwaves and walls, and it must retransmit anything that arrives damaged. A connection that delivers 90 percent of advertised speed over a cable often holds 60 percent over Wi-Fi one room away. Ethernet is not glamorous, but it is the single cheapest upgrade for anyone whose downloads matter.

Sharing comes next. Your advertised speed describes the pipe coming into the building, not a private lane to your device. A household streaming two screens, syncing cloud folders and updating phones simultaneously divides the same capacity, and your single download shows only its share. Background traffic on the same machine - telemetry, sync clients, automatic updates - takes a smaller but real bite of the same budget.

Then the server side, which no plan upgrade can fix. A download is half-guided by the host: overloaded mirrors, rate-limited free tiers, distant servers behind congested routes all cap the pace regardless of your line. Finally, a subtle unit footnote: operating systems sometimes report file sizes in binary units, where a MiB runs about 4.9 percent larger than a decimal MB, adding a small apparent discrepancy. None of these factors is theft. They are the standing costs of a real network, and knowing them by name is what separates diagnosis from grumbling.

CHAPTER 05Reading Speed Numbers Without Getting Fooled

Speed tests and download clients report different quantities, and comparing them directly is a category error. A speed test measures your line's current capacity against a nearby, well-provisioned server - the best case for your connection. A download client reports an actual transfer from one specific source, with its own limits and route. A healthy setup shows the client figure between 60 and 100 percent of the test figure; a client figure far below that range points at the source or the wireless link, not the plan.

Watch the units on every screen. Speed tests speak in Mbps to match your bill; download clients and file managers usually speak in MB/s to match your files. The ratio to remember is 8:1, so 100 Mbps should appear as roughly 12.5 MB/s before overhead and 9 to 11 MB/s in ordinary conditions. If a client shows kilobytes instead of megabytes, or bits instead of bytes, the apparent gap can be pure unit confusion - check the abbreviation before concluding anything is wrong.

The same literacy applies to shopping. Comparing a 300 Mbps plan to a 500 Mbps plan is comparing 37.5 MB/s to 62.5 MB/s of ceiling - and the improvement you feel depends on whether your actual bottleneck is the line, the Wi-Fi, or the servers you download from. If big files dominate your evenings, run the arithmetic on your typical file sizes before paying for capacity that overhead and Wi-Fi will quietly tax. The Bandwidth Calculator's custom speed field makes that a two-minute exercise with numbers you can trust.

🔑 Key takeaways

• Mbps (bits) and MB/s (bytes) differ by exactly 8: divide bits by 8 to get bytes, so 100 Mbps equals 12.5 MB/s of ceiling.
• Advertised speed describes raw capacity; real transfers commonly run at 60-90% of it, so expect download times about 1.1x to 1.7x the ideal.
• Protocol overhead, Wi-Fi losses, shared devices and server limits are the named causes of the gap - each one is checkable.
• Speed tests measure your line's best case; download clients measure one real transfer. Comparing them directly misleads.
• Per-hour math converts speeds into volume: a 100 Mbps line moves about 45 GB in an ideal hour.
• Before upgrading a plan, confirm the bottleneck is actually bandwidth - the fix for Wi-Fi losses or a slow server is not a faster line.

❓ Frequently asked questions

📚 More in the Toolfyra blog · or browse all free online tools.