What It Actually Costs to Store Your Audio (and Why the Billing Looks the Way It Does)

In short:

• Two billing shapes cover almost everything you pay for: usage-based (pay for exactly what you use) and prepaid capacity (pay a flat fee for a bucket you may or may not fill).
• The math almost always favors usage-based, for anyone who doesn’t reliably fill their tier — which real data says is most people.
• Flat rates persist anyway, because of well-documented psychology, not because they’re the better deal for most customers.
• The gap between what you pay for and what you use has a name — breakage — and it’s a real, deliberate, multi-billion-dollar line item for a lot of companies.
• own.audio charges a flat $0.05 per GB per month, on purpose, specifically to not build breakage into the design — the reasoning is at the end, transparently.

The two shapes billing comes in

Almost every recurring bill you pay is one of two shapes, whatever industry it’s in.

Usage-based billing charges you for what you actually consumed, measured after the fact. Your electricity bill. Your water bill. A pay-per-visit gym pass. A phone plan billed per minute and per megabyte, the way most mobile billing worked before “unlimited” plans existed.

Prepaid capacity billing charges you a flat fee upfront for a bucket of capacity — a tier — regardless of how much of it you actually use. A gym membership. A 2 TB cloud storage plan. An all-you-can-eat buffet. An “unlimited” phone plan. Audible’s one-credit-a-month subscription, where a credit costs the same whether it buys a 4-hour audiobook or a 40-hour one.

Both shapes are legitimate ways to sell something. Neither is a scam. But they behave completely differently once you actually use them, and the difference is almost never explained plainly — which is the point of this guide.

One distinction worth making early, because the two get blurred constantly: strictly, “prepaid” means money handed over before use, sitting as a balance that depletes — a gift card, an Audible credit, a phone top-up. A Dropbox or iCloud+ tier is technically a recurring flat fee, billed monthly regardless of usage, not a balance that runs down. The two aren’t identical mechanically. But they produce the exact same economic shape for the customer — a fixed price for a capacity ceiling, paid whether or not it’s reached — which is why this guide treats them as one category with the same underlying math and the same psychology, and why a term like “prepaid capacity” is used loosely throughout to mean both.

What the math actually says

Start with the plainest version: cloud storage, because the numbers are public and the comparison is direct.

Three of the biggest consumer storage providers sell 2 TB as a flat monthly tier:

Provider	2 TB tier	Per-GB equivalent
iCloud+	$9.99/month	$0.0049/GB
Google One	$8.33/month (year one; check current terms — some plans step up after the first year)	$0.0041/GB
Dropbox Plus	$11.99/month	$0.0059/GB

Those per-GB rates look tiny — and they are, if you’re actually using close to 2 TB. Almost nobody storing audio is. A large audiobook collection runs tens to a couple hundred gigabytes; even a serious ripped-CD music library rarely reaches a full terabyte. Buy 2 TB to store 80 GB of audiobooks, and your real per-GB rate on a $9.99/month iCloud+ plan isn’t $0.0049 — it’s:

$9.99 / month ÷ 80 GB actually stored = $0.125 / GB / month

That’s 25 times the tier’s advertised per-GB rate, because the other 1,920 GB you’re paying for sits empty. This is true of every flat tier, for every customer below full capacity — the advertised per-GB price is only real at 100% utilization, and utilization is rarely anywhere close.

There’s a second layer to this worth knowing before shopping for a tier: smaller tiers are almost always priced worse per gigabyte than larger ones, from the same provider. iCloud+ is a clean example — its published tiers, at the per-GB rate they imply:

iCloud+ tier	Monthly price	Implied per-GB rate
50 GB	$0.99	$0.0198/GB
200 GB	$2.99	$0.0150/GB
2 TB	$9.99	$0.0049/GB

The 50 GB tier costs four times as much per gigabyte as the 2 TB tier — a deliberate structure, not an accident. It nudges anyone doing the math toward the bigger tier, which is exactly where the gap between what’s bought and what’s used tends to be largest. A household that would genuinely be served by 80 GB is priced, gigabyte for gigabyte, into looking at 2 TB instead.

Now compare that same 80 GB library against a flat usage-based rate. At own.audio’s $0.05/GB/month:

80 GB × $0.05/GB = $4.00/month

Less than half of the cheapest 2 TB tier above, for the amount of audio a real household actually stores — not because usage-based pricing is a trick, but because a tier is sized for a hypothetical maximum, and billed at that maximum’s price whether you reach it or not.

Where the crossover actually sits

This isn’t a rule that usage-based pricing always wins — it depends entirely on how much of a tier you actually fill. Using the cheapest tier above (iCloud+, $9.99 for 2,048 GB) against a flat $0.05/GB rate, the crossover point is:

$9.99 ÷ $0.05/GB = 199.8 GB

Below roughly 200 GB of real usage, the per-GB rate is cheaper. Above it, the flat tier is. The honest, complete answer to “which is cheaper” is: work out your actual library size and compare it to that number — not a general rule that either model always wins.

Worked out at a few realistic library sizes, against the cheapest 2 TB tier compared above, the pattern holds consistently:

Actual library size	Usage-based ($0.05/GB)	Cheapest 2 TB flat tier ($9.99)	Cheaper option
20 GB (a modest audiobook collection)	$1.00/month	$9.99/month	Usage-based, ~90% less
80 GB (a large audiobook collection)	$4.00/month	$9.99/month	Usage-based, ~60% less
150 GB (audiobooks plus a modest music library)	$7.50/month	$9.99/month	Usage-based, ~25% less
200 GB (right at the crossover point)	$10.00/month	$9.99/month	Effectively tied
500 GB (a large, long-collected music library)	$25.00/month	$9.99/month	The flat tier, ~60% less

Most audio libraries — audiobooks especially, which run smaller than most people assume — land well to the left of that crossover point, which is why the tiers built for a maximum keep costing more than the audio inside them would cost priced on its own.

Why so many people pick the flat rate anyway

If the math this often favors usage-based pricing, why do flat tiers dominate the market? Because the decision isn’t made on math — it’s made on a well-documented set of psychological effects, and the research behind them is worth taking seriously rather than dismissing as “people are bad at math.”

The foundational study here is Stefano DellaVigna and Ulrike Malmendier’s 2006 paper in the American Economic Review, built on real attendance data from three US health clubs and 7,752 members over three years. The finding: members who chose a flat monthly membership over $70 attended an average of 4.3 times a month — paying more than $17 per visit — when a 10-visit pass at the same clubs cost $10 per visit. Averaged across their membership, those members forgo roughly $600 in savings they could have had by simply buying visit passes instead. They weren’t confused about the price; they were wrong about their own future behavior, systematically and predictably.

That same year, Anja Lambrecht and Bernd Skiera published a parallel finding in the Journal of Marketing Research, naming the pattern flat-rate bias: choosing a flat fee over pay-per-use even when pay-per-use is objectively cheaper. Their research traced it to four separate psychological effects, and it’s worth naming all four, because different ones apply to different products:

• The insurance effect — paying extra for the certainty of a fixed bill, the same reason people buy insurance against the odds. You’re not paying for the audio storage itself; you’re paying to never have to think about a surprise charge.
• The taxi-meter effect — a running, visible cost causes real discomfort even when the total is smaller than a flat fee would be. Watching a number climb feels worse than not watching a number at all, independent of which one is actually higher.
• The overestimation effect — people systematically predict they’ll use more than they do. This is the direct mechanism behind the gym study above: members didn’t want a 4.3-visit-a-month membership, they intended to go far more often, and priced their commitment against that intention rather than their eventual behavior.
• The convenience effect — a flat fee removes a small decision every time you’d otherwise have to weigh “is this worth the marginal cost,” which is a real, if small, source of daily mental friction that a lot of people are willing to pay to avoid.

The insurance effect deserves a moment on its own, because it’s the one case where paying more than the expected cost is a fully rational choice, not a bias — that’s the entire business model of insurance itself. Nobody buys car insurance expecting to total their car; they buy it so that if the unlikely happens, the cost is a fixed premium instead of a bill that could ruin them. Flat-rate pricing borrows exactly that logic for far smaller stakes: a fixed $10 monthly storage bill is a worse bet than usage-based pricing on average, but it is a guaranteed number, and for someone who values not having to think about a bill over saving a few dollars a month, that’s a legitimate trade — not a mistake.

Current mobile data plans show the same pattern at a much larger scale. One widely cited US study found that 56% of people on unlimited data plans use less than 10 GB a month, with the average person in that group using just 5.1 GB — and that switching to a capped plan sized to actual usage could save the typical over-provisioned unlimited subscriber $268.44 a year. A separate UK study put the average unused allowance at 3.4 GB per subscriber per month — 143 million GB going unused across the market every single month. Different country, different providers, same underlying behavior: people buy more capacity than they use, reliably and at scale, because the alternative — a bill that could vary — feels riskier even when it’s usually smaller.

Mobile carriers themselves know this well enough to have a name for the opposite failure mode: “bill shock” — the jarring experience of a metered bill coming in far higher than expected. T-Mobile maintains an official consumer page dedicated to helping customers avoid it, with automatic overage alerts built into the network. That a carrier spends real engineering effort warning customers away from usage-based surprises is itself evidence for the insurance effect above — the anxiety of a variable bill is treated as a real problem worth solving, independent of whether the variable bill is usually smaller than a flat one would have been.

It isn’t just tech — the same pattern runs through completely unrelated industries

If this were only true of cloud storage and phone plans, it would be a quirk of one market. It isn’t. The identical shape — a flat fee for capacity most people don’t fill — shows up in industries with nothing else in common.

Airline miles are, structurally, prepaid capacity for future travel, and the breakage is enormous. Industry estimates put roughly 35 trillion frequent-flyer miles sitting unredeemed globally, representing about $1.75 trillion in lost travel value and generating an estimated $200 billion a year in breakage profit for the airlines and their partner banks. US airlines alone reported a combined $27.5 billion in unused loyalty balances in 2020 — American Airlines carried $9.2 billion of that on its own, Delta $7.2 billion, United $6 billion. Typical breakage rates across loyalty programs run 10–20% of issued points, depending on the program — a fifth of what’s promised, on average, never gets claimed.

Subscriptions in general show the same waste even without a “tier” in the traditional sense. CNET’s most recent annual subscription survey found Americans wasting roughly $252 a year on subscriptions they don’t use, with 59.9% of people carrying at least one unused paid subscription in a given month — an average of 2.6 unused subscriptions each. Nobody sets out to pay for something they won’t use; the flat-fee structure just makes it easy not to notice until a survey asks.

Gym memberships were the specific product studied in the founding research above, and the mechanism generalizes cleanly to anything sold as “unlimited access, billed monthly, whether or not you show up” — a category that also includes streaming subscriptions, meal-kit deliveries, and software seats nobody logs into.

The pattern is consistent enough across genuinely unrelated industries — travel, telecom, fitness, cloud storage, subscriptions of every kind — that it’s worth treating as a structural feature of how flat-fee pricing works everywhere, not a coincidence specific to any one of them.

A practical way to work out which is actually cheaper for you

The math above generalizes into a formula anyone can run on their own numbers, for storage or almost anything else sold both ways:

1. Track your real usage for a typical month (not a guess — check an app, a bill, a settings screen)
2. Divide the flat tier's price by that real usage number
3. Compare the result to the usage-based rate

If step 2’s number is higher than the usage-based rate, the flat tier is currently costing more than pay-as-you-go would — regardless of how cheap the tier’s advertised per-unit rate looks on the label, because that label rate only applies at full capacity. For audio storage specifically: check how large your actual library is (most apps show this in settings, or it’s visible as a folder size), divide any flat tier’s price by that number in GB, and compare the result to $0.05. Below the crossover point worked out earlier — roughly 200 GB against the cheapest 2 TB tier compared here — the flat tier is very likely priced above what a matching usage-based rate would charge for the same audio.

Why it works just as well for the people selling it

The psychology above explains why customers choose flat tiers. It doesn’t fully explain why companies love selling them — and that side of the story has its own name: breakage.

Breakage is the accounting term for revenue a company gets to keep from capacity or credit a customer paid for and never used. It is not a loophole or a dark pattern — it’s a standard, disclosed line item in a lot of companies’ financial reporting, and the scale of it is genuinely large. Americans currently hold somewhere in the range of $21–23 billion in unredeemed gift card balances, money that was paid for and never spent. Starbucks alone reported $212 million in breakage revenue in a single year (2022), out of $1.77 billion in outstanding unredeemed card balances — a real company, disclosing a real number, for exactly this mechanism.

Storage tiers and subscription plans work the same way structurally, even without an explicit “breakage” line in the annual report: a 2 TB tier sold to someone storing 80 GB is, functionally, the company being paid for 1,920 GB it never has to actually deliver. This is also part of why flat tiers can be priced so aggressively low per advertised gigabyte — the provider isn’t pricing against 100% utilization, it’s pricing against the realistic average, which is far below the tier’s ceiling.

Audible’s credit system is a version of this closer to home for this site’s audience. One credit a month, usable for one audiobook of any length or price — meaning a credit spent on a 6-hour book is worth far less than one spent on a 30-hour book, and a credit not spent at all rolls over, up to a limit of six, before expiring twelve months after it was issued. It’s a well-liked, well-designed system as these things go — but it’s still, structurally, prepaid capacity: money committed today against audio you may or may not actually redeem it for within the window given.

Spotify’s audiobook allowance is a stricter version of the same idea, and a cleaner illustration of breakage in action: Premium subscribers get 15 hours of audiobook listening a month from the subscriber catalog, included in the price — and per Spotify’s own support documentation, any hours not used by the end of the month are simply gone, with no rollover at all. A listener who uses 6 of their 15 hours in a slow month isn’t carrying 9 hours forward; that unused capacity resets to zero and Spotify keeps selling the same 15-hour allowance to the next subscriber, every month, regardless of how much of it any individual actually claims.

None of this makes flat-rate or credit-based pricing dishonest. It makes it a real, working business model, chosen deliberately, that happens to be more profitable on average than usage-based billing precisely because of the psychology in the section above — not despite it.

“Paying for the air”

Put the last two sections together and there’s a plain way to describe what’s actually happening: with prepaid capacity, you’re paying for the difference between what you bought and what you use — the empty space, the unused minutes, the credit that expires — as much as you’re paying for the part you actually consume. It isn’t a hidden fee. It’s the entire model, and it’s the reason the same 2 TB storage tier can look inexpensive per gigabyte on the label and expensive per gigabyte in practice.

That’s not automatically the wrong choice for every person or every product — the insurance effect and the convenience effect are real, legitimate reasons some people are happier paying more for certainty. But it’s worth being able to name the tradeoff plainly before choosing it, rather than only seeing the advertised per-GB number on the label.

When a flat rate genuinely is the better choice

None of the above means usage-based pricing wins for everyone — it means it wins for anyone who doesn’t reliably use most of what they’d prepay for, and it’s worth being specific about the real exceptions rather than implying there aren’t any.

Heavy, predictable users come out ahead on tiers. The DellaVigna and Malmendier gym data cuts both ways: members who actually did attend enough to make the flat fee the better deal existed in their dataset too, just as a minority. Anyone whose audio library genuinely sits near or above a tier’s ceiling, consistently, is in the segment a tier is priced for — the crossover math earlier in this guide is exactly how to check which side of that line an individual actually falls on.

Budget certainty has real value that isn’t irrational. A family on a tight, fixed monthly budget may reasonably prefer knowing the audio bill will never be a cent more than $10, even if a usage-based bill would average $7 most months — the insurance effect again, and a legitimate preference, not a mistake to be talked out of.

Simplicity has a real price worth paying, for some people. Not tracking a running balance, not checking a dashboard, not thinking about it between bills — for someone who would find that mental overhead genuinely unpleasant, a flat fee buying silence is a fair trade even at a premium.

The honest version of this guide’s argument isn’t “usage-based pricing is always right.” It’s “usage-based pricing is right far more often than the popularity of flat tiers suggests, and the reason for the mismatch is psychology, not math” — which is a different, more useful claim, and one that still leaves room for a flat fee to be the correct choice for a specific person’s specific situation.

Why own.audio charges per GB instead

This is the part where it’s fair to say plainly: this reasoning is also why own.audio’s own pricing works the way it does, and it’s worth being transparent about, rather than just stating the number.

The model: a flat $0.05 per GB per month, no tiers, no minimum commitment. A family storing 200 GB of audiobooks, music, and podcasts pays:

200 GB × $0.05/GB = $10.00/month

Store less, pay less, automatically — no bucket sized for a maximum you’re not using.

How billing actually works day to day: rather than a recurring subscription charge, own.audio uses a credit system — you add credit (say, $10–20), and it’s drawn down gradually based on how much you actually have stored. The app always shows the current daily cost, the current monthly cost, and the exact date the credit will run out at the current rate. Auto-recharge is optional, not the default.

What happens if credit runs out: playback pauses, but nothing is deleted on the spot. The library stays in place for a real grace period, giving time to add credit or download files, with deletion only after an extended period of continued inactivity — not the moment a balance hits zero.

That grace period isn’t just a courtesy — it’s built around how people actually treat audio they’ve collected. Families tend to keep audiobooks and recordings rather than delete them, even ones they’ve already finished, the same way a bookshelf doesn’t get emptied after a book is read. A billing model that could quietly delete a library over a lapsed payment would be actively hostile to that behavior; a grace period respects it instead.

Why this shape, specifically: it’s the version of billing that doesn’t have breakage built into it by construction. There’s no tier sized above what most people use, no unused capacity quietly generating margin, no credit expiring on a clock chosen by the platform rather than the customer. It’s less profitable per customer, on average, than a well-designed flat tier would be — that’s the direct tradeoff for not pricing against unused capacity — and it’s the one chosen anyway, because it’s the version where the bill and the actual usage are the same number.

Is it the cheapest possible option in every scenario? No — nothing is, at every library size, against every competitor’s promotional pricing. Families with genuinely large libraries who have compared real bills should judge the tradeoff for their own case, the same way this whole guide has argued anyone should. What per-GB pricing does guarantee, by construction, is that the bill and the actual usage are the same number — not a bet on a ceiling that may or may not get used.

Whether that stays the only option is genuinely open. If real usage data ever shows enough people who’d be better served by an optional flat tier alongside the credit model — not instead of it — that’s a fair thing to revisit. Nothing like that exists today, and this article isn’t laying groundwork for one; it’s an honest explanation of the choice already made, and the reasoning stays valid regardless of what, if anything, gets added to it later.