soda Journal
You shouldn't have to gamble on a travel data plan
Short version: don’t pick a travel eSIM by the number of gigabytes. Pick it by how much you’re willing to spend. You can’t predict your data, but the budget is yours to set.
The night before you fly, you open the checkout page. Three gigs, five, ten, unlimited. And you try to work out a number you have no way of working out: how much will I actually use? You can’t. And that isn’t your fault.
”How many gigs do you need?” is the wrong question
Nobody can predict their own data use before a trip. Whether the hotel Wi-Fi holds up, whether a change of plans puts you on maps all day, whether a friend keeps sending videos: you don’t know any of it yet. Data is something you only understand after you’ve used it. A traditional eSIM asks you to commit before then.
Look at it another way and you’re being asked to price your own behaviour a week from now. You’re not buying data, you’re placing a bet on how much you’ll use, and the bet is rigged so either direction costs you.
So one of two things tends to happen. Bet too big, and you buy 10GB but use 4GB; the rest expires and you paid for nothing. Bet too small, and you run dry on day four, then top up abroad at a worse short-term rate.
Either way, you’re the one who pays more. The seller does fine both times. A good chunk of this industry runs on you guessing wrong.
A better question: count dollars, not gigabytes
The question isn’t “how many GB should I buy.” It’s “how much am I willing to spend to stay online this trip.”
| Counting GB (traditional) | Counting dollars (soda) | |
|---|---|---|
| Before you leave | Predict usage, pick a plan size | Pick nothing |
| Use too little | Leftover data expires | You only pay for what you used |
| Use too much | Buy another bundle, or pay overage | Auto-capped at the cheapest plan price |
| How it feels | Worrying the whole trip | Land, open it, go |
| Who pays for a wrong guess | You | No one, because there’s nothing to guess |
Counting gigabytes makes prediction your job and bills you when you get it wrong. Counting dollars drops the prediction, since it was never something you could get right.
Here’s the actual math
Say you take a five-day trip to Tokyo and use 2.8GB.
On a fixed plan, you bought 5GB, used 2.8, and the leftover 2.2GB expires. The bill still says 5GB.
With soda, we settle the 2.8GB you used, and we check the price for you. If the 3GB plan price comes out cheaper than metering 2.8GB unit by unit, we cap you at the 3GB plan price and charge whichever is lower.
That’s all “capping” is. You’re billed by usage, but never more than a plan would have cost. Use a little and you pay a little. Use a lot, and the ceiling is the cheapest plan price.
(Figures here are illustrative. Real per-country prices land at launch.)
Two travellers, one card: two more scenarios
That 2.8GB was the middle case. Real trips skew further than that, so let’s run the two ends and you can place yourself.
The light user. A couple spends four days in Kyoto, mostly temples, cafes, and walking. The hotel has Wi-Fi, so photos go home in the evening. They might burn just over 1GB the whole trip. On a fixed plan they’d probably buy 3GB to be safe, use a third of it, and watch the rest expire. With soda they pay for roughly that 1GB and the bill is a fraction of everyone else’s.
The heavy user. One person spends seven days in Osaka, streaming on the bullet train, livestreaming to friends while walking around, and hotspotting a travel buddy who has no data. They clear 2GB a day without trying. On a fixed plan they’d reach for 15GB or an unlimited bundle and pay a big sum up front. With soda they use a lot, but the bill auto-caps at the cheapest plan price, so heavy use doesn’t get punished.
Here’s the point: neither person had to sort themselves into a tier before leaving, and they don’t buy different products. The same soda card charges the light user little and caps the heavy user. You don’t need to know in advance which one you are, and that’s exactly what makes counting dollars the calmer way to travel.
| Traveller | Actual use (illustrative) | What a fixed plan does | What soda does |
|---|---|---|---|
| Kyoto, 4 days, light | ~1GB | Buy 3GB, 2GB expires | Pay for ~1GB |
| Osaka, 7 days, heavy | ~15GB | Buy unlimited, pay a lot up front | Cap at cheapest plan price |
| Typical 5-day trip | ~2.8GB | Buy 5GB, 2.2GB expires | Lower of actual use or plan price |
(Usage and amounts are illustrative; real figures land at launch.)
Usage-based sounds unpredictable? The cap is your safety net
A lot of people flinch at “pay by usage,” because telco metered billing has historically been a bottomless pit: go over and the meter just keeps running. soda isn’t that. Ours is metered billing with a ceiling. You get the upside of paying little when you use little, without the risk of an unbounded bill when you use a lot. The cap is your worst case, and it equals the price of a plan you could have bought anyway, never worse.
So what you actually get is the flexibility of pay-as-you-go plus the price guarantee of a fixed plan. Both upsides, neither downside.
Why bundles were the only option before, and aren’t now
Prepaid bundles made sense once. Early billing systems couldn’t settle accurately by real usage, so “sell a pack, cut off when it’s empty” was a fair, technically simple compromise. But real-time rating has matured. Systems can now meter as you go and pick the cheapest math at checkout. When the technology can do that and a seller still makes you bet on a bundle, it comes down to one reason: it pays better.
A bundle pays the seller on three levels. The cash lands first. Anything you leave unused is close to pure margin. And once you’ve bought, you stop shopping around. Tidy for them, and each layer is a small loss for you. soda would rather not. We don’t want to profit when you guess wrong.
Honestly, we’re not the cheapest
If you’re after the lowest sticker price on the market, soda probably isn’t your pick. We run on proper carrier lines and treat a stable connection as the floor, not the cheapest re-routed traffic we can find. Out on the road, a map that won’t load or a payment code that keeps failing costs you far more than the dollar you saved.
We’re not chasing cheapest. We’re chasing the version where you don’t lose, which is a different goal. Cheap usually bills you somewhere you can’t see: throttled speeds, weaker routing, data that expires unused. Not losing means the money you pay lines up with what you got, and you can read every line of it.
Common questions
So how do I pick a soda plan? You don’t. A traditional eSIM makes you choose a country and a GB size first; soda removes that step. Land, open it, and you’re billed by actual usage, auto-capped at the cheapest plan price. The only thing you decide is the budget you’re comfortable with, and that ceiling is the cap.
Could pay-as-you-go end up costing more than a plan? No. The cap is a ceiling: you keep the savings when you use little, but you never pay more than the cheapest plan would have cost. Your worst case equals a plan you could have bought anyway, never worse. The mechanics are in overage and auto-renew.
Does unused data expire? soda doesn’t sell a “bundle,” so there’s no expiry and your balance never zeroes out. The most frustrating part of fixed plans, paying for data you never use, is gone by design. More in leftover data.
Do I still need to look up “how much data for 5 days in Japan”? Look it up if you want a ballpark, but you no longer have to turn that number into a purchase. For an estimate, see the activity table in how much data do you need; to buy, soda spares you committing to a number at all.
soda isn’t the cheapest, so where’s the value? The value isn’t the lowest sticker price, it’s that you don’t lose whichever way the trip goes. Cheap cards hide the cost in throttling, re-routed lines, and data that expires unused; soda runs proper lines, never expires your balance, and shows every line of the bill. See the hidden cost of cheap eSIMs.
Staying online abroad shouldn’t make you gamble before you’ve left the house. Leave the worry at home, and let soda handle the math.