International operations · 2026

Cold calling multiple countries: geo permissions, real rates, local numbers and calling hours

Last updated September 3, 2026

Adding a second country to an outbound motion looks like a list problem and turns out to be an operations problem. The first call fails for a reason nobody warns you about, the budget you built from a single "per minute" figure is wrong by several multiples, half the local numbers you wanted cannot be bought without an address you do not have, and the local caller ID you were counting on may be blocked at the far end. This page is the checklist, with rates read off Twilio's own country pricing pages in September 2026 rather than borrowed from a comparison post.

1. Geographic permissions: the call that fails before it rings

Twilio classifies every country's number ranges into low risk and high risk. Low-risk ranges — ordinary fixed and mobile numbers — are reachable from an account out of the box. High-risk ranges are not: they cover premium, shared-cost and special-service numbers, plus narrow ranges Twilio associates with international revenue share fraud, and Twilio refreshes those lists several times a month. Trial accounts are restricted to low-risk destinations entirely.

So the usual story is not "international calling is off" but something more annoying: most of a new country works, one number in twenty does not, and nobody connects the two events. Add the two other common causes — an administrator who tightened permissions earlier for fraud reasons, and an account still in trial — and geographic permissions become the single most common reason a first international campaign stalls.

The fix takes a minute: Console → Voice → Settings → Geo permissions, find the country, enable it. If a specific number still fails after the country is on, enable the high-risk categories one at a time rather than all at once, so you keep the fraud protection that the setting exists to provide. Permissions are set per project, so a subaccount you spun up for a new market does not inherit them — a genuinely common source of "it worked yesterday". Before any new market: enable the country, place one test call to a colleague there, and only then import the list.

2. Why one "per minute" number is meaningless

Domestic US pricing has trained everyone to think of a call as costing about a cent and a half. That intuition survives Canada and dies everywhere else. Twilio Programmable Voice list prices, taken from each country's own pricing page, checked September 2026, in USD:

The last column is one rep, all-in for a month: 1,029 billed minutes to that country's mobiles at 120 dials a day, plus one local number, plus $29 of DialSheet Pro, plus the $0.0040 per-minute browser leg every WebRTC call carries in addition to the leg out to the prospect. Same rep, same effort, same software: $49 in the US against $132 in Ireland. That is the whole argument for costing a market before you staff it.

Three patterns worth carrying forward. Fixed lines are cheaper than mobiles everywhere except the US and Canada, where the North American numbering plan makes them identical — so a switchboard-heavy enterprise list is structurally cheaper to dial in Europe than an SMB list of direct mobiles. Number rentals are noise against minutes at any real volume, with one exception: a German mobile number is $30 a month against $1.35 for a German local one. And every rate above is a list price before Twilio's automatic volume tiers, which start mattering long after most sales teams stop reading.

3. Numbers: who wants proof of a local address

Wanting a local number and being allowed to hold one are different things. Twilio collects country requirements through Regulatory Bundles, and the requirements are set by the local regulator, not by Twilio, which is why they vary so much. What we verified on Twilio's regulatory guidelines in September 2026, for a business buying a voice number:

Two rules generalise. First, a PO box or virtual address is not accepted as proof of a local address anywhere — the one workaround people always reach for is the one that is explicitly excluded. Second, the national and mobile ranges are consistently laxer than geographic ranges, so when a market blocks you on the local number, check whether a national number gets you in. One trap in the other direction: French mobile numbers are restricted to person-to-person use and prohibit automated outbound calling, so they are the wrong number to put a dialer behind even if you can get one. Requirements change — Australia's incoming Scam Prevention Framework is expected to tighten its checks — so read the country page before you plan around any of this.

4. Local presence across a border, and where it now stops

Local presence works: a prospect answers a number that looks like a neighbour more often than one that looks like an overseas call centre. Across a border, three limits apply, and only the first is the one people expect.

• You must be entitled to the number. Presenting a number you hold and can be called back on is local presence. Presenting one you do not hold is spoofing, and it is what every scam-call regime in this list was written to stop. The line is ownership, not intent.
• Some countries now block foreign-origin domestic caller ID outright. Since 29 January 2025 Ofcom expects UK providers to block calls that arrive from abroad presenting a UK caller ID, outside a narrow set of exceptions — including calls a provider can demonstrate are made by a UK customer even where the traffic originated on a network outside the UK. Whether your UK-CLI call from Toronto or Dublin lands therefore depends on how your provider originates and attests it. Australia has had an industry code on reducing scam calls since 2022 with its own outbound CLI obligations, and European regulators increasingly treat a domestic caller ID arriving over an international route as high risk by default. Ask your provider before you commit a campaign to a foreign presentation number.
• Attestation does not travel. STIR/SHAKEN signing is a North American scheme, so an A-attested US number tells a UK or German carrier nothing. Reputation has to be built per market, which is an argument for a small stable pool of numbers per country rather than a large rotating one. Our attestation guide and local presence guide cover the mechanics.

5. Calling hours are a scheduling problem, not a policy one

Every regime measures hours in the recipient's local time, so a multi-country list needs to be segmented by country and, inside large countries, by region. The windows are not the same shape, and the strictest one governs whichever segment you are dialling:

Two practical consequences. Australia and Canada both lose whole days or half-days at weekends, so a team selling into both plus the US has a very uneven week, and Australian public holidays are state-specific. Separately, the suppression list has to be global: a person who asks to be removed in one market must not reappear when a different rep imports a different list. Screening obligations remain per country — the National DNCL in Canada, TPS and CTPS in the UK, the Do Not Call Register in Australia, the NDD in Ireland, the federal and state registries in the US — and none of them substitutes for another. See DNC and TCPA basics for the US position.

6. How per-seat dialers price international minutes

This is where multi-country teams find the gap between a quote and an invoice. Per-seat tools — Aircall at ~$30–50/user/mo, Kixie at ~$95–185/user/mo (quote-only), PhoneBurner at ~$140–183/user/mo (annual), list prices as of August 2026 and approximate — bundle or meter minutes inside a licence that is priced per user. When an allowance is described as unlimited or generous, it is normally unlimited within the vendor's home domestic footprint; international destinations are metered separately, at rates that are rarely on the public pricing page.

That is not a scandal, it is a wholesale pass-through with a margin on it. But the margin is applied exactly where the underlying cost is highest, so the countries that hurt most at wholesale hurt most on a seat too — and you cannot see the number in advance. The question to ask is never the seat price; it is "what do you charge per minute to an Irish mobile, and does it count against the bundle?" A vendor who answers in one line is fine to buy from. The same trap in its domestic form is dissected in what "unlimited minutes" really means.

The bring-your-own-carrier answer is structurally different rather than merely cheaper: you read the carrier's published rate for the destination and pay the carrier that rate, so there is nothing to discover on the invoice. One account holds numbers in as many countries as you buy them in, and the software price does not move with geography — $29 per 3 seats, free for one person up to 500 calls a month. The trade is that you own the setup: geo permissions, bundles and number reputation are yours to manage, which is the work this page describes. DialSheet also offers a managed line that needs no carrier account, but it is not available outside the US and Canada, so for a genuinely multi-country team the answer is your own Twilio.

The checklist for adding a country

1. Read the destination's per-minute rate for both fixed and mobile, and re-cost the campaign at the mobile rate.
2. Enable the country in Voice geographic permissions — in the right project — and place one test call.
3. Check the number requirements before promising anyone local presence; if the local range needs an address you do not have, price the national or mobile range instead.
4. Ask your provider whether a foreign-origin call presenting that country's CLI will be delivered.
5. Register for, or subscribe to, that country's do-not-call screening, and add its output to one global suppression list.
6. Set the calling window in the recipient's time zone and split the list by region.
7. Warm the new number rather than opening at full volume — see warming up a new number.

Questions people actually ask