A video customer who loses money is not a retention tool

Most regional cable operators now run a video business they would never start today. Programming costs rise every year, customers leave every month, and the set-top boxes that deliver the service depreciate on the balance sheet long after anyone wants them. Yet video bundles survive, defended by an argument that sounds like finance: video customers are less likely to cancel their broadband, so the video loss is the price of keeping more profitable customers. I believed that argument for a long time. When we measured it at Ridgeburn, it turned out to be mostly a description of who our video customers were.

The bundle argument, tested

The raw figures are persuasive. In 2022 our customers who took both video and broadband churned at about a third of the monthly rate of broadband-only customers. Our chief revenue officer, who had spent his career in cable sales, saw that difference as the main reason to keep promoting bundles, especially with fixed-wireless competitors arriving in our towns.

The difficulty is that bundled customers differ from broadband-only customers in ways that have nothing to do with the bundle. Ours were older on average, had been with us longer, owned their homes more often and paid by automatic debit more often. Every one of those characteristics predicts lower churn on its own. Comparing the two groups directly attributes all of that difference to video.

We built matched cohorts instead, pairing bundled and broadband-only customers with similar tenure, town, speed tier, payment method and billing history. Most of the gap disappeared. The bundle itself was associated with about a quarter of a percentage point less monthly churn. That is not nothing. Spread over our base, it was worth a few million dollars a year in avoided losses. But our fully allocated video business was losing about $4 million a year, and the loss was growing by roughly the amount programming costs rose. We were paying more for retention than the retention was worth, and the gap widened every renewal.

Allocation matters to this argument, and it is where many analyses go wrong. If programming is the only cost charged to video, the product looks profitable, because revenue exceeds what we pay the programmers. But video customers also generate far more care calls and truck rolls than broadband customers, they need boxes that we buy and depreciate, and they consume a share of billing and collection. Charging video for all of that is not punitive. It is simply the account a board needs to decide anything.

Retransmission from both sides of the table

Part of the rising cost is retransmission consent, the fees local broadcast stations charge distributors to carry their signals. I know this market from both sides. At Merrimont Broadcasting Group, from 2008, I built the per-subscriber model the company used to negotiate with about forty cable and satellite distributors. Those fees rose from $19 million to $86 million in five years, and they rescued a station group whose advertising had collapsed. The stations had a fair case: their local news and network programming were among the most watched channels any distributor carried, and for years they had been paid nothing for them. Part of each increase also went straight to the networks the stations were affiliated with, which expected a share of the fees.

At Ridgeburn I am on the paying side, at about $14 a video customer a month for local stations alone. Renewals run on cycles of roughly three years, and a station group negotiating with a small regional operator has more market power than we do, since its other distributors are far larger. When agreements lapse without a deal, a station's signal can go dark on our system, and our customers blame us rather than the station.

Having sat across the table helps in a particular way. I can show a broadcaster's negotiators that we understand their obligations and their arithmetic, and then show them ours: how many customers we lose for each dollar we add to the bill. Since 2023 we have shown local broadcast fees as a separate line on customers' bills, so that the cost of a renewal is visible to the people paying it. That has made our renewals more honest on both sides, if not cheaper.

Running a decline on purpose

In 2023 the board approved four changes. We stopped buying set-top boxes and began moving customers to an app-based service on streaming devices they already owned. We ended promotional video bundles for new customers. We declined to renew a regional sports network that charged $5.40 a month for every basic customer, though fewer than one in ten video households watched it for more than an hour a month. And we wrote off $9 million of box inventory immediately and shortened the remaining lives of boxes in customers' homes, rather than carrying them at a value the business no longer supported.

Before changing any package, we read every carriage agreement for minimum-penetration and tiering clauses, because a programmer can require that its channel reach a set share of customers, and moving it to a smaller tier can trigger higher rates rather than savings. Two contracts had to be renegotiated before we could act.

The decline was not painless. About 4,200 customers left within six months of the sports network's removal, and we estimate 1,700 of them also moved their broadband. Some of the town councils that grant our franchises heard from angry residents before they heard from us, which was our failure. Video customers fell from 129,000 at the end of 2022 to 84,000 at the end of 2025.

The account changed as intended. Video contribution moved from a $4 million loss in 2022 to an $11 million profit in 2025, and about $17 million a year of box spending moved into fibre. I do not argue that every operator should leave video, or that bundles never retain anyone. My argument is narrower: a board should see the fully allocated video account, a retention effect measured on matched customers rather than raw averages, and the clauses that decide what a package change saves. Then it can decide how fast to let the business decline, which is a strategic choice. Keeping it unchanged because of an untested belief is not.