Bandwidth for Online Course Platforms: How Much You Need

Every instructor comparing platforms looks at monthly price and student caps. Almost nobody looks at bandwidth, which is the number that actually decides whether a plan fits — and the one that produces unpleasant surprises in month four, when the course finally sells.

Storage and bandwidth are not the same thing

They get conflated constantly, and they behave completely differently.

Storage is the space your video files occupy. You pay for it once, it does not change when people watch, and it is small. Storage is almost never the constraint.

Bandwidth is the data transferred every time a student plays a video. It multiplies by every viewer, every rewatch, every time someone scrubs back to hear a sentence again.

The distinction matters because plans advertise both, and instructors instinctively worry about the wrong one. When a plan lists a modest storage figure and a large bandwidth figure, the bandwidth number is doing all the work.

The arithmetic

You do not need a table. One number is enough:

Monthly bandwidth ≈ hours of video watched per month × 0.7 GB

That figure assumes course video recorded sensibly. Most course content is a person talking, slides, or a screen recording — very little motion, which compresses far better than generic video advice assumes. Record it badly and the same hour can cost you two or three times as much. The recording guide covers the settings.

The key word in the formula is watched, not enrolled.

And expect launches to be the case that breaks a plan, not steady state. A few hundred students enrolling in one week and working through the opening lessons can consume most of a month's allowance in days. Size for the spike, not the average.

Why "unlimited bandwidth" is never unlimited

Delivering video costs the provider real money. There is no pricing model where that stops being true, so when a platform offers unlimited bandwidth, one of three things is happening.

It is capped in the fair-use clause. The marketing page says unlimited; the terms reserve the right to throttle or review accounts with "excessive" use, undefined. You find the real limit when you hit it.

It is funded by commission. Your bandwidth is genuinely uncapped because the platform takes a share of every sale. The heaviest users are also the highest-revenue users, so it balances. A legitimate model — just understand you are paying for bandwidth in proportion to your success, on a different line.

It is throttled quietly. Playback quality drops for your students when you become expensive. You do not get a bill; you get complaints about buffering, which is worse.

None of these are scandals. What matters is knowing which one you signed up for, because it changes what happens on your best month rather than your average one.

A platform that states its figures plainly, and lets you raise a limit without a full upgrade, is easier to plan around than one that says "unlimited" and defines it later.

Four ways to cut the bill

1. Record for what your video actually is

Talking-head and screen-recorded content has very little motion. Recording it with settings intended for general video wastes a large share of every gigabyte. This one change is usually the biggest saving available and costs nothing but a re-export.

2. Let quality adapt, and do not default to the top

A good platform offers several quality levels and lets each player choose. What matters is the default: starting everyone at the highest rung burns your maximum on a student watching in a phone-sized window where they cannot tell the difference. Starting sensibly and stepping up on demand is a large, invisible saving.

3. Trim dead air

Ten hours of lessons routinely contain half an hour of pauses, restarts and "let me just find that tab". Every second of it is delivered to every student. Editing is worth doing for the learner's sake; the bandwidth saving compounds with every enrolment.

4. Do not put static slides in video

A twenty-minute video of one unchanging slide with a voiceover costs the same bandwidth as twenty minutes of full-motion video, and teaches less than a downloadable document would. Where the visual genuinely does not move, audio plus a document is a better lesson at a fraction of the cost.

What counts against your allowance

Plans rarely spell this out, and the gaps are where estimates go wrong.

Counts, and people forget: rewatching. Students revisit the hard lesson three times, and scrubbing backwards pulls the same data again. A student who "completed" a five-hour course may have used far more than five hours of data.

Counts, and people really forget: your own testing. Reviewing uploads, checking playback on four devices, walking a colleague through the platform — all of it consumes what a student would.

Counts: free preview lessons. A public preview is served to everyone who lands on the page, including people who never buy. A popular preview can quietly become your single largest consumer.

Usually does not count: images, documents, page loads. Real, but far smaller than video. Do not model them.

Depends on the platform: downloads for offline viewing. Some count the download, some count it once and never again, some do not offer it. If offline access matters to your students, ask specifically.

Bandwidth and student caps are one constraint in two hats

Plans usually pair a student limit with a bandwidth limit, meant to fit each other for a typical library. Which one you hit first tells you something useful.

Bandwidth first? Your courses are long, or your students are unusually engaged. That is a good problem — high completion is what produces testimonials and referrals — but it means you should size on bandwidth and treat the student cap as incidental.

Student cap first? Your catalogue is short or completion is low. Worth investigating: if hundreds enrol and few watch, the constraint on your business is not your plan.

Watch the ratio rather than either number alone. It tells you what to fix.

A low number is not always good news

If your consumption sits well under your estimate and students mention buffering, you are probably not being efficient — you are being served slowly. Players drop quality when video arrives slowly, which lowers your bandwidth while making the experience worse.

So test playback the way a student experiences it: on a phone, on an ordinary mobile connection, not on your office network. And ask a platform where video is served from for your audience's region before you commit.

Choosing a plan without guessing

Do this before you sign up, not after:

1. Count your hours of finished video, or estimate what you will have in six months.
2. Estimate students per month at the point you would call the business working.
3. Assume well under half complete the course, not all of them.
4. Multiply: hours × students × completion × 0.7 GB.
5. Add headroom for a launch spike and rewatching.

Then compare that against the plan, and ask what happens when you exceed it.

If the arithmetic puts you far under an entry plan, you have a year before this matters — build content instead of worrying about it. If it puts you near the ceiling, that is not a reason to buy the biggest plan today. It is a reason to check that raising the limit later is possible without a migration.

Where your video is hosted also changes these economics, sometimes dramatically. The hosting comparison covers what each option actually costs once your course is selling.