How much does roaming actually cost your company? The method, in four buckets

Almost no company can answer this question, and the reason is not carelessness: roaming spend arrives in four different shapes, on four different lines of the invoice, and nobody owns all four. Travel owns the trips, finance owns the invoice, IT owns the contract, and the charge that hurt was on a page none of them reads. The number you want is boring and singular: cost per traveller per trip. You can get it out of your own invoices in an afternoon, and once you have it, every decision about roaming becomes arithmetic instead of opinion.

This page is the method. It does not tell you what to buy. It tells you what you are paying now, which is the part that is usually missing when someone proposes a change.

Why nobody knows the number today

Three ordinary things hide it.

The spend is split across cost centres. Per trip passes appear on the mobile invoice. A pass an employee bought themselves appears in expenses. A local SIM bought at an airport appears as a receipt with no category. The same cost lands in three systems that are never added together.

The spend is mostly small. Nine charges out of ten are the size nobody escalates. They are visible only in aggregate, and aggregating them is work that has no owner.

And the spend is counted per line, not per trip. Carrier billing is organised by SIM and by month, because that is what a contract is. Travel is organised by trip and by person. Nothing in the invoice tells you that those two charges in March were the same trip to Riyadh.

If you have already had the loud version of this problem, the causes are catalogued separately in roaming bill shock on a work trip. This page is about the quiet version, which is larger.

Step 1: pull the last twelve months of mobile invoices

Twelve months, not three. Business travel is seasonal, and a quarter that happens to contain a sales conference or a summer shutdown will give you a number you cannot plan with.

Ask the person who holds the carrier relationship for the itemised version, not the summary. What you need from each invoice:

• the recurring charges per line, which is where a monthly international add-on sits, month after month, whether anyone travelled or not;
• the one off charges, which is where per trip passes and day passes sit;
• the usage charges outside the bundle, which is the per megabyte or per day billing that applies when no pass was in place;
• the line identity, so a charge can be attached to a person.

Add to that the same twelve months of expense claims filtered for anything telecom shaped: airport SIM cards, hotel wifi, a personal eSIM purchase, a data top up. That stack is usually thin and always instructive, because it is where the traveller went around a process that did not work.

Step 2: sort every charge into four buckets

Four buckets, because each one is caused by a different thing and responds to a different lever. Mixing them is what makes the total look unfixable.

Two rules that keep the exercise honest.

Bucket 4 comes out of bucket 2, not in addition to it. A shock invoice is out of bundle usage; you separate it so that it does not smear across the average and make normal trips look expensive. Pick a threshold before you start, for example any single trip that generated more than 150 in charges, and apply it consistently.

Bucket 3 is the one to count in months, not in money. For each line carrying a monthly add-on, write down how many months that person actually left the country. That column is the finding.

Step 3: count travellers and trips

Two counts, and both are usually easier than they sound.

Travellers is not headcount. It is the number of distinct people who went abroad on company business in the twelve months. A company of forty often has twelve.

Trips is the number of separate journeys outside whatever zone your contract includes at no extra charge. If travel is booked centrally, the flight booking system has this. If it is not, expense reports have it, because a trip that generated no expense claim almost certainly generated no roaming either. Count a two country trip as one trip and note the countries; you will want that later when you look at which destinations actually cost you money.

While you are there, record nights per trip. The daily pass buckets scale with nights, and a four night average against a nine night average changes which option wins.

Step 4: divide, and get the only number you need

total roaming spend (buckets 1 to 4)  /  trips  =  cost per traveller per trip

Also keep the same total divided by travellers, which is the annual cost of equipping one travelling employee, and the four bucket subtotals, which is where the levers are. Everything else is decoration.

A worked example, with the assumptions in the open

These numbers are round and invented, chosen so the arithmetic is easy to follow and easy to redo with your own. They are not our prices, and they are not anyone's prices. The example is in one currency; use yours, the method does not change.

Assumptions

• A company of 40 employees, of whom 12 travel abroad on business.
• 60 trips in twelve months to destinations outside the zone their contract covers, averaging 4 nights.
• 35 of those 60 trips had a pass bought for them, at 10 per day for 4 days.
• 25 trips had no pass, and generated an average of 24 in out of bundle usage each.
• 6 lines carry a monthly international add-on at 12.50 per month, all year.
• Of those 6 people, 2 travel most months and 4 travelled twice in the year.
• 2 invoices in the year contained a large one off charge: 400 and 300.

The arithmetic

The numbers that come out of it

• Cost per traveller per trip: 3,600 / 60 = 60.
• Cost per travelling employee per year: 3,600 / 12 = 300.
• Trips per traveller per year: 60 / 12 = 5.

And the finding that only appears once bucket 3 is counted in months

The 4 people who travelled twice paid the add-on for 12 months each: 4 x 12 x 12.50 = 600. The months they actually travelled account for 4 x 2 x 12.50 = 100. The remaining 500 bought nothing at all. It is the clearest piece of waste in the example, it sits in a recurring charge that nobody ever decided to keep, and it is invisible in any view that shows roaming as a single number.

Redo those six lines with your own invoices and you will have a better roaming report than most companies of your size.

What each bucket responds to

Now the useful part, which is that the four buckets do not respond to the same thing. Treating them as one problem is why roaming projects stall.

Bucket 1, passes per trip, responds to buying the right size once. Day passes are priced per day and renew per day, so a long trip pays the daily rate for every day of it. EE Business, for example, lists a Business Daily Passport at "£9.32 each day" with 1GB and a World Daily Passport at "£11.64 each day" with 1GB (EE Business, roaming, read 30 September 2026, no date shown on the page). AT&T's International Day Pass is "$12/day" for the primary line and "$6 for each additional line used on the same calendar day" (AT&T, International Day Pass, read 30 September 2026, no date shown on the page). Neither of those is expensive for two nights. Both of them are arithmetic you should do before a nine night trip, and before a trip where two colleagues travel together.

Bucket 2, out of bundle usage, responds to one setting and one sentence in a policy. The setting is data roaming, off on the company line unless a pass is in place, switched per line in the phone's own settings. The sentence is which zone your contract covers, written where travellers will see it, because "roaming included" and "roaming included in Europe" are read identically by someone packing at midnight. EE states the boundary plainly for its business plans: "No EE Business plan includes mobile data outside our Europe Zone as standard" (same page, read 30 September 2026). Also worth knowing which trips generate this bucket: it is almost always a laptop, and how much data you need for a work trip has the per day figures that explain why.

Bucket 3, monthly add-ons, responds to the months column and nothing else. Keep the add-on for the two people who travel most months, where it is the cheapest thing available. Stop it for the four who travel twice, where it is a subscription to nothing for ten months of the year. This is the one lever that requires no behaviour change from any traveller, which is why it is usually the first one to pull.

Bucket 4, shock invoices, does not respond to price at all. It responds to a limit that exists before the spend does. A cheaper per megabyte rate applied to a laptop syncing overnight is still a number nobody approved. What changes bucket 4 is a rule that refuses at the point of purchase, or a plan bought in advance with a fixed amount, so the ceiling is known before the trip rather than after it.

What a company account changes, and what it does not

For completeness, here is what our own company account does, because the shape of it matters more than the price when you are comparing against the four buckets above.

The company pays, the employee uses. The employee buys in the Wigoo app on their own phone, the eSIM is theirs, and the bill goes to the company. One eSIM installs on one phone. The employee signs in with their work email and a six digit code, on a verified company domain or a listed address, and that verification lasts ninety days.

The company sets rules the server checks at every purchase. A monthly cap per person, a cap per purchase, the largest plan allowed, which destinations the company pays for, and whether top ups are covered. A purchase outside the rules is refused, with the reason shown in the app before the button rather than on an invoice next month, and the employee can still buy it with their own card if they need it. That is the mechanism that acts on bucket 4.

The company sees spend, destinations and who. It does not see what the employee browses, and by design it does not see their data consumption. Expenses export as CSV.

Three limits worth knowing before you plan around it. It is prepaid balance only, with no invoices yet. A company account is opened by us, not by self service signup. And our plans are data only: no phone number, no calls, no SMS. The employee's own line stays in the phone, so the work number and the SMS codes keep working, which is the part people worry about and the part that does not change. If "business eSIM" means something else to you, the two products that share the name are separated in business eSIM explained, and if your phones are managed by IT, whether an eSIM can go on them at all is answered in can I add an eSIM to my work phone.

What a plan costs for the destinations you actually travel to is a conversation, not a table: tell us where your people go and we will talk it through.

Sources: EE Business, "Roaming", read 30 September 2026, no date shown on the page, quoted for the Business Daily Passport and World Daily Passport day rates and for the statement about data outside the Europe Zone. AT&T, "International Day Pass", read 30 September 2026, no date shown on the page, quoted for the day pass rate per line. Carrier rates differ by country, by contract, by tariff generation and by date, and business price lists are commonly quoted excluding VAT while consumer ones include it, so compare like with like against your own price list rather than against a public page. Every figure in the worked example is illustrative and rounded for legibility: it is not a Wigoo price, not a carrier price, and not a measurement.

Quick answers

We do not log trips anywhere. How do I get a trip count?

Use expense claims as the proxy. A trip abroad that generated no expense claim almost certainly generated no roaming either, so the claims give you a count that is wrong in the safe direction. Flight bookings are better if travel is booked centrally, and calendars are a last resort that works surprisingly well for a single team. You need the order of magnitude, not a perfect number: the difference between 60 trips and 65 does not change any decision, and the difference between 60 and 200 changes all of them.

Which bucket should we act on first?

The monthly add-ons paid in months when nobody travelled, because it is the only lever that needs no traveller to do anything differently. You cancel a recurring charge for the people whose months column says two, keep it for the people whose column says ten, and nothing else about how your company travels has to change. Out of bundle usage is usually second, because it is one setting per line.

Does a company account mean our people stop using their work line abroad?

No, and it would be a bad idea if it did. Our plans are data only, with no phone number, no calls and no SMS, so the employee's own line stays in the phone and keeps the work number and the SMS verification codes working. What moves to the travel profile is the data, which is the part that generated the charges you were counting.