This article makes one claim, and it is arithmetic, not a survey result: the real price of a fixed data package is the package price divided by the data you actually use. Use 75% of a package and you have paid 33% more per gigabyte than the sticker said. Use half and you have paid double. Nobody measured that. It is division, and it applies to every fixed bucket of data ever sold.
We sell metered data, so we have an obvious interest in this argument, which is exactly why every number below is either our published rate or a calculation you can redo on a napkin. You will find no industry overpay statistic in this piece, because we do not have a source we trust for one and we are not going to invent it. What we can show is the mechanism, and why it stays invisible at the only moment it matters: the moment you pay.
The arithmetic
Two lines cover the whole subject:
- Effective price per GB = package price / gigabytes actually used
- Overpay vs sticker = (1 / utilization) - 1
The sticker price per gigabyte is what the shelf shows. The effective price is what you actually paid once the trip is over and the unused remainder has expired. The gap between the two depends on a single variable, utilization, and it does not grow linearly:
| Share of the package used | Overpay vs sticker per-GB price | A $10 / 5 GB pack: effective per-GB |
|---|---|---|
| 90% (4.5 GB) | +11% | $2.22 |
| 75% (3.75 GB) | +33% | $2.67 |
| 60% (3 GB) | +67% | $3.33 |
| 50% (2.5 GB) | +100% | $4.00 |
| 40% (2 GB) | +150% | $5.00 |
Note the shape. Missing full utilization by a little costs a little. Missing it by a lot costs disproportionately more, because the same fixed price is being spread over fewer and fewer gigabytes. And the order of events is the entire problem: you find out which row you were in only after the trip, but you pay before it.
Worked example: Spain, ten days
Take a ten-day trip at roughly 300 MB per day, an ordinary mid-range figure for maps, messaging, social apps and photos: about 3 GB for the whole trip. A natural package size to buy against that estimate is 5 GB, because nobody wants to run out. Call it a $10 pack, which is a $2.00 per gigabyte sticker. Those are neutral round numbers chosen for arithmetic, not a quote of anyone's price list.
The trip goes exactly as planned and you use your 3 GB. Utilization is 60%, so your effective price is $10 / 3 GB = $3.33 per gigabyte, 67% above the sticker that convinced you. On metered billing at our Spain rate of $1.43 per GB, the same 3 GB costs $4.30 in total, and the effective per-gigabyte price is $1.43 no matter how the trip turns out, because there is no unused remainder to spread the cost over.
An honest footnote: at this particular sticker the package loses even at 100% utilization, since $2.00 is above our Spain rate and a full 5 GB metered is $7.17. That is a property of cheap-rate countries, not a general law. The next example runs the other way.
Worked example: Japan, where the package can win
Japan meters at $3.38 per GB with us, one of the pricier rates on our list. Bulk packages can sticker below that, so suppose one at $20 for 10 GB, or $2.00 per gigabyte, well under our rate. If you reliably use all 10 GB, the package simply wins: $20 against $33.79 metered, and nothing in this article argues otherwise.
Now run the same ten-day, 3 GB trip through it. Metered: $10.14. The package: still $20, with an effective price of $6.67 per gigabyte, 233% over the sticker that made it look like the obvious choice.
The crossover is one division away: $20 divided by $3.38 per GB is about 5.9 GB. Use more than that and the package was the right buy. Use less and metered was. The package's problem is not its price. It is that the purchase forces you to answer the 5.9 GB question before you leave, and a wrong answer in either direction is paid for by you, never by the seller.
Underbuying: the second package
Overbuying is only half the trap. Suppose you go the other way on the Spain trip: buy a 3 GB pack, and a couple of heavier days mean you run out on day eight. The fix is genuinely unremarkable. You buy another 3 GB over hotel wifi in two minutes, at the same price as the first one. Nobody price-gouges you, and there is no drama to report.
The arithmetic is the quiet part. You use around 400 MB in the remaining two days and fly home with 2.6 GB of the second pack untouched. The whole trip in one line: 6 GB bought, 3.4 GB used, 57% blended utilization, 76% over sticker. At the $2.00 sticker that is $12 paid for 3.4 GB, an effective $3.53 per gigabyte. Underbuying does not escape the utilization trap. It re-enters it, because a pack bought two days before you fly home has almost no trip left to be used in.
Why you cannot simply estimate better
The obvious counter is to size the package correctly, and for travellers with stable habits that can work. For most people, per-day usage genuinely resists prediction, for reasons that have nothing to do with discipline:
- Navigation scales with being lost. Route recalculations, satellite layers and transit lookups in an unfamiliar city can turn a 5 MB maps day into a 50 MB one, and the only thing you notice is that you found the restaurant.
- Video quality is chosen by the app, not by you. The same ten minutes of short-form video costs several times more when an app decides your connection deserves 1080p instead of 480p. Defaults shift per network, per app version, per day.
- One hotspot hour can outweigh a phone day. A laptop behind your phone pulls OS updates, cloud sync and mail attachments that a phone would defer. A single working session can use more than the rest of the trip's phone traffic combined.
- Background sync picks its own moment. Photo backup waits for a good connection and then takes it, wherever that connection happens to be metered.
Our guide to how much data you actually need gives per-activity ranges, and they are honest, wide ranges: an ordinary sightseeing day lands anywhere from 200 to 600 MB, while a quiet resort day can sit under 150 MB. A spread that wide is precisely what makes committing to one bucket size in advance a coin flip.
When a package is the right call
Symmetry demands this section, and it is not a grudging one. A fixed package is the better purchase when three things line up: the sticker per-gigabyte price is below the metered rate for your destination, your usage is high and steady enough to land near full utilization, and the validity window comfortably outlasts the trip. A two-week stay in Japan where you hotspot a laptop every working day clears all three.
The general rule falls straight out of the arithmetic: the package wins whenever your eventual utilization ends up above the ratio of sticker price to metered rate. In the Japan example that ratio is $2.00 over $3.38, about 59%. If you can honestly predict you will land above it, buy the package. The trap is not that packages are overpriced. It is that this prediction is the one thing the checkout page cannot help you with.
What metered billing changes
Per-megabyte billing does not make data cheaper by the gigabyte. In some countries our metered rate sits above a bulk package's sticker, as Japan just showed. What it changes is structural: you pay after usage instead of predicting it, so utilization is 100% by construction and the effective price always equals the sticker price. There is nothing to size, nothing to expire, and no late-trip second purchase that goes home mostly unused.
For calibration, our current rates run from $0.92 per GB in France, $1.43 in Spain and $1.54 in Turkey to $3.28 in the United States and $3.38 in Japan; the full list for all 193 countries is on the prices page. The balance itself never expires, which is its own subject and covered in the no-expiry piece, and the mechanics of how per-MB metering actually charges are in the pay-per-megabyte explainer.
None of this makes packages a scam. A package is a bet on your own predictability, sold without an odds table. The tables above are the odds. Find your row before you pay, not after.
FAQ
How much do you overpay on a data package you don't fully use?
It is arithmetic, not a statistic: divide the package price by the gigabytes you actually used. Using 75% of a package means you paid 33% over the sticker per-gigabyte price, 60% used means 67% over, half used means you paid double, and 40% used means 150% over. A $10 pack of 5 GB with 3 GB used works out to $3.33 per GB against a $2.00 sticker. The percentage follows from your own utilization, so place yourself in the band rather than trusting any quoted average.
How much data do I need per day of travel?
Honest ranges, not precision: light use (messaging, maps, email, some photos) usually lands around 100 to 300 MB per day. An ordinary sightseeing day with social apps and short video runs about 300 to 600 MB, while a quiet resort day can sit under 150 MB. Streaming or hotspotting a laptop can push a single day past 1 to 2 GB, and one long hotspot session can exceed an otherwise normal week. That wide spread is exactly why sizing a fixed package in advance is hard.
Is pay-as-you-go always cheaper than a package?
No. Metered billing with Roamzy in Japan costs $3.38 per GB, and a bulk package can sticker lower; the worked example in the article uses $2.00 per GB. If you reliably use the whole package, the package wins: $20 for 10 GB beats $33.79 metered. The crossover in that example is about 5.9 GB. Metered wins when usage is below the crossover or genuinely unpredictable, because its effective price always equals its sticker price.