The nine-day customer.

A business whose customers are all leaving

Call the archetype nine days. It could be four or it could be three weeks — the exact number varies by market, season and route, and anyone quoting a single global figure is guessing. What matters is the shape: a customer who arrives, transacts once, uses the product continuously for a short defined window, and then disappears on a date known in advance.

A mobile operator would describe that as catastrophic churn. In travel connectivity it is not a problem to be solved. It is the specification.

Churn is not a leak in this business. It is the product working correctly.

The inversion

Once you accept that, most of the received wisdom of the telecom industry stops applying — and a surprising number of companies have lost money discovering this the expensive way.

What breaks when lifetime value goes away

Consumer telecom is, financially, a lending business dressed as a service. An operator spends heavily to acquire a subscriber — subsidised handsets, commissions, marketing — and recovers that spend slowly across many months of billing. The entire model is an amortisation schedule.

Remove the months and the schedule collapses. There is no second month to recover anything in. Whatever it costs to acquire and serve this customer must be recovered inside a single transaction, or it is never recovered at all.

Read that column again and notice what it implies. If there is no retention, no upsell and no loyalty, then the only lever left is the cost of doing the transaction at all. This is not a marketing business. It is an operations business wearing a consumer face.

Where the risk actually sits

Plot the nine days honestly and they are not flat. Revenue lands entirely at the start. Cost and risk pile up in a narrow spike immediately after — and then, if everything worked, almost nothing happens for a week.

This shape has a hard consequence. You are paid in advance for a promise, and the promise is tested in a foreign country by someone who has just got off a plane. If it fails there, you refund, you absorb a support cost that may exceed the sale, and you lose the channel partner who recommended you. All of the downside is compressed into the first hour.

Bought at the worst possible time

Consider the conditions under which the purchase decision is actually made. The traveller is typically:

• Under time pressure
  In an arrivals hall, or minutes before boarding. The window for conversion is measured in minutes, sometimes seconds.
• Poorly informed
  Unsure what a gigabyte means for their nine days, which networks are good locally, or whether their phone even supports what they are buying.
• Mildly anxious
  Connectivity is the precondition for the map, the ride, the hotel confirmation and the person waiting. It is not a luxury purchase; it is a relief purchase.
• Willing to overpay to stop thinking
  Which is precisely why the airport kiosk and the roaming pack survive despite being poor value. They are bought to end the problem, not to solve it well.

Any product that adds a step at this moment loses. Not because the step is difficult, but because the customer's tolerance for steps is close to zero. The winning experience is not the cheapest or the most featured — it is the one that finishes fastest.

The customer is acquired by someone else

Here is the part that reorganises the whole business. Given a decision window of minutes and near-zero brand loyalty, the seller is whoever happens to be standing next to the traveller at that moment.

And that is almost never the mobile network. It is the travel agent who booked the trip, the booking confirmation email, the tour operator, the hotel front desk, the bike rental counter, the visa consultancy. These businesses already hold the traveller's attention, already have their trust, and already know their destination and dates.

In this business, distribution is not how you reach the customer. Distribution is the customer relationship.

Why channel beats brand

Which explains something that otherwise looks strange: enormous, well-capitalised mobile operators with excellent networks routinely lose this market to small intermediaries with no infrastructure at all. The intermediary is not winning on network quality. They are winning on proximity to the decision.

Why cost-to-serve is the entire game

Put the pieces together. Small basket. Single transaction. No repeat purchase to average across. Margin measured per unit, not per relationship. Volume is the only route to a real business.

In that arithmetic, any human touch is fatal. A support ticket, a manual activation, a call to reconcile an invoice, a person checking a passport — each of these can cost more than the sale that triggered it. In a subscription business, one support call across a two-year relationship is a rounding error. Here it can wipe out the transaction entirely.

This is why automation in travel connectivity is not an efficiency programme. It is the business model itself. And it is also why the metrics that matter are not the telecom ones:

1. Activation success rateWhat percentage of purchases result in a working connection without any human intervention. The single most important number in the business.
2. Time from purchase to first byteMeasured in seconds. This is the customer's entire experience of quality.
3. Support contacts per thousand activationsA direct read on cost-to-serve, and the earliest warning that something upstream is broken.
4. Refund and failure rate by routeFailures cluster by destination, device and network. Aggregate numbers hide the pattern that would let you fix it.
5. Inventory utilisationProfiles are generated in advance. Unsold stock is a real cost; stockouts are lost sales at the exact moment demand peaked.

Why incumbents are structurally bad at this

It is tempting to assume large operators ignore this market out of complacency. The truer explanation is structural, and it is worth understanding if you intend to compete with them.

Their billing systems were built for monthly cycles and identified account holders. Their support organisations are staffed for a subscriber base, not a burst of arrivals in high season. Their commercial teams are measured on subscriber additions, and a customer who leaves after nine days does not count as one. Their inter-operator settlement runs on cycles of weeks, against a customer whose entire lifetime is shorter than the settlement period.

Each of those is a rational design for the business they are actually in. Together they make the nine-day customer look, from inside a large operator, like noise — a small, awkward, seasonal revenue line that fits none of the existing machinery.

Markets like that do not stay unserved because nobody noticed. They stay unserved because serving them requires a different machine.

What the nine days actually demand

Read back through the constraints and they converge on a single, specific set of requirements. This business needs capacity that can be bought in small increments and priced per unit rather than per contract. It needs provisioning that completes in seconds without a human in the loop. It needs settlement fast enough that a nine-day customer does not outlive the payment. And it needs to reach the thousands of small businesses that stand next to travellers at the moment of decision — none of whom will ever sign a bilateral agreement with a mobile network.

That list is not a product. It is an infrastructure specification, and it describes a layer that mostly does not exist yet. The technical pieces are all in place: profiles can be delivered anywhere in seconds, and capacity is abundant. What is missing is the wholesale machinery between them — the part that lets a network's supply reach a counter in another country, priced, provisioned and settled, without anyone picking up a phone.

The nine-day customer is not a niche. It is a rehearsal. Any market where demand is short, cross-border and intermediated by thousands of small sellers ends up needing the same layer. Travel connectivity is simply where the gap is most visible — and where the customer is kind enough to tell you the exact date they will leave.