Since 1 July 2019, leaving a UK mobile network has taken one text message. You text the word PAC to 65075, and your provider has one minute to text back the code that lets you take your number with you. Hand the code to a new network and the switch must finish within one working day. No call to the retention desk, no exit interview. A regulator did that on purpose.
In the first year of that rule, customers requested almost 5.8 million of those codes. About 4.1 million became real switches. The remaining 1.7 million are the most interesting number I have come across in weeks: people who asked for the code, held it for up to thirty days, and never used it.
What the counter caught
Ofcom, the UK communications regulator, called the reform text-to-switch, and it had teeth. The code must arrive within a minute. The new provider gets one working day to complete the switch. Providers were banned from billing notice periods that run past the switch date, which Ofcom estimated would save customers 10 million pounds a year. In the twelve months before the reform, 4.4 million codes had been requested and 3.5 million used. In the twelve months after, requests jumped by about 1.4 million while completed switches rose only about 600,000.
Read those two jumps against each other, because the gap is the lesson. Actual leavers rose by about a sixth. Requests rose by nearly a third. Most of the new codes were not exits at all, and Ofcom's 2023 evaluation offers the likely explanations: people checking whether they were still under contract, and people requesting a code as leverage to renegotiate with the network they already had. A code is a switch only when it is used. An unused code is a raised hand. The code cost nothing to request, so asking became a cheap way to make your dissatisfaction legible to the only party that could price it.
Here is the opinion this piece exists for. Every audience I have ever worked with can produce its arrival number instantly, because an arrival is an event - a visit, a signup, a click, a timestamp. Almost none can produce an exit number, because leaving fires no event anywhere. It is the absence of a next visit, and absence does not trigger anything. So the arrival count gets printed on the homepage and celebrated in meetings while the drain runs unmeasured under the totals - even though the drain, not the inflow, decides whether those totals are growth or replacement.
Friction never kept anyone. It hid the leaving, and a hidden leaving reads as devotion on every dashboard you own. UK telecoms is the one industry where a regulator forced the exit to become a measurable, timestamped event, and the measurement's first discovery was 1.7 million people who wanted a better deal from the network they already had. No arrival counter on earth could have produced that number. It is dissatisfaction, priced continuously.
The honest caveats, because they matter. Carriers did not collapse: completed switches rose to 4.2 million in 2021. Ofcom's economists could not confidently establish how much of the increase the reform itself caused. And the consumer benefit Ofcom calculated - 130 million pounds over a decade, above its own 2017 forecast - was driven partly by take-up: around 71 percent of customers requested their code through the automated routes, against the 51 percent Ofcom had predicted. Making the exit easy did not just count it. It got used.
The bill for staying
The other half of the reform attacked what the hidden exit was quietly earning. Since 15 February 2020, UK phone, broadband and pay-TV providers must warn customers between 10 and 40 days before their contract ends: what they have been paying, what they will pay once it ends, and the provider's best deals, including prices offered only to new customers. Anyone already out of contract must be reminded once a year.
The reason was in Ofcom's research from the same period. Around 20 million UK households were sitting past their initial contract, and staying put without renegotiating cost about 20 percent more for customers bundling landline with broadband, and about 26 percent - 10 pounds a month - for those who stacked pay-TV on top. Ofcom called it the loyalty price premium. The exit nobody measured was never free. The people who never left were quietly paying for it.
Counting the way out here
The applied lesson is not to count who comes back - that is the arrival counter grown up, and this blog has already argued for it. It is: give the exit a shape you can count, then read the raised hands. Your version of the switching code is any visible, one-step way out - a pause option, a downgrade tier, a cancel page with a reason box - wired to a counter. Count how many people reach for it, and how many take the alternative you hand them on the way out. That pair of numbers is the closest most businesses ever get to knowing who was about to leave, and the ones you keep are your 1.7 million.
Counting arrivals while never counting exits is running a company that logs every hire and no resignation.
I will be honest about my own house. The counter on our homepage reads 3,660 visitors since October 2, as I write this - a proud arrival number, published the way everyone publishes theirs, and it cannot tell me one thing about who stopped coming. The board underneath it, though, is built so that its own exits are countable. The November race closes on the 30th, a countdown in days and hours printed on the front page. Rank is your total bid value for the month, ties broken by the most recent payment, and when the month turns over the board resets with the minimum bid back to a dollar. When a slot lapses, the board does not hide it: an empty rank is a counted exit on public display, the exact thing most audiences never produce. No loyalty premium can pool here, because no position outlives the month that bought it.
A slot buys arrivals - introductions in front of a browsing crowd, every placement labeled - and whether those arrivals stay is the half of your books no leaderboard can keep. Count who reaches for the way out first. Then buy the arrivals: as I write this the #1 spot is open at a dollar, at bidspotlight.com/submit.