Every subscription business runs the same promotion eventually. Twelve months at half price, and the acquisition chart goes vertical. I bought media for DirecTV during the years when that playbook was the entire category, and the number nobody put on the launch slide was the one that decided whether any of it made money: what happens in month thirteen, when the price steps up and the subscriber remembers they were never really a customer. They were a coupon with an account number.
The month-thirteen cliff is not a surprise to anyone who has watched a promo cohort mature. It is a schedule. Subscribers acquired on a deep discount churn at multiples of the rate of full-price subscribers, and they churn precisely when the discount expires, which means the acquisition team's best quarter and the retention team's worst quarter are the same event, separated by a year, reported by two departments that do not share a dashboard. The acquisition team got its bonus in Q1. The churn shows up the following Q1, attributed to retention, pricing, content, anything but the coupon that caused it. I have sat in the meetings where both teams presented on the same afternoon, each with green numbers, describing one broken machine from opposite ends.
Cohorts or fiction
The only honest way to read a subscription business is by cohort, and most reporting refuses, because blended numbers are kinder. Blended churn averages the loyal base against the promo tourists and produces a figure that alarms nobody. Split it by acquisition offer and acquisition channel, and the business snaps into focus: full-price subscribers from brand search behaving like annuities, deep-discount subscribers from comparison sites behaving like a scheduled refund. Run the lifetime math on both and the ranking of your acquisition channels usually inverts. The channel with the cheapest cost per subscriber is routinely the most expensive cost per retained subscriber, and only one of those numbers is a business.
Put real figures on it. A subscriber acquired at $80 who pays a $40 promotional rate for twelve months and leaves at the step-up delivered $480 of revenue against acquisition cost, content cost, and service cost. A subscriber acquired at $200 who pays full freight and stays three years delivered several times that at margin. The first one won the weekly report every single week of her tenure. The second one won the only report that matters, and in most subscription companies that report has never been built, because building it requires the acquisition offer to sit in the same table as the month-thirteen outcome, and no one who owns the offer volunteers for that join.
Discounts are targeting
Here is the part that never makes it into the promo postmortem: the discount does not just attract price-sensitive customers, it manufactures them. An offer is a filter, the same way a headline is. Lead with half off and you have selected, at scale, for the exact population most likely to leave at full price, then handed that population to a retention team and asked why loyalty is down. The offer taught them what the service is worth, and they believed it. The subscription brands that age well treat the intro offer as an audience decision, not a conversion lever: shallower discounts, shorter bridges, offers built around commitment rather than price, and creative that sells the thing rather than the markdown, because whoever converts on the product survives the step-up and whoever converts on the coupon was always leaving on schedule.
The objection arrives on cue: without the aggressive offer, acquisition volume drops, and the growth chart the board sees goes flat. True, and worth saying back slowly. The volume being defended is volume that mostly refunds itself in month thirteen. A subscription business that cuts promo depth and watches gross adds fall while twelve-month retention climbs has not slowed down. It has stopped renting subscribers and started keeping them, and the P&L notices before the board deck does. The transition quarter is uncomfortable. Every quarter after it is cheaper.
The operating rule I took from those years is one sentence: no acquisition offer gets approved without its month-thirteen retention forecast attached, from the cohort history of the last offer like it. If the forecast does not exist, the history has not been read, and the promotion is a guess with a budget. Subscription economics forgive almost everything except pretending the calendar is not coming. It comes every month, thirteen at a time.