Subscription marketing budgets, sized by what a subscriber is worth.

The rest of this page covers how to do that in practice, why the average cost per subscriber points to the wrong budget, how to tell channels that bring valuable subscribers from channels that bring many, and what a model for a recurring-revenue business has to get right.

Ask a subscription business how much it should spend on marketing, and the answer is usually last year's budget plus or minus a bit, or a share of revenue. Both ignore the number subscription businesses have that most others lack: a fairly stable estimate of what a new customer is worth.

Lifetime value turns the budget question into a break-even question. If an average new subscriber brings a certain gross margin over their lifetime, that margin is the most you can pay to win one and still break even. Most businesses want payback inside a set horizon, often twelve or eighteen months, so the useful figure is the margin a subscriber brings within that horizon. That figure is your ceiling for what one new subscriber may cost.

From lifetime value to a budget, in three steps

1. Work out what a subscriber is worth on margin. Take average revenue per subscriber over your payback horizon and multiply by gross margin. Revenue overstates what you can afford to pay.
2. Make that the ceiling for the next subscriber. This is the most the next subscriber may cost, in any channel. Many teams set the ceiling a little lower to leave room for overhead and uncertainty.
3. Find the budget where the next subscriber reaches the ceiling. Below that level, more spend wins subscribers who pay back. Above it, each extra subscriber costs more than they bring in.

The third step is where most teams get stuck, because it needs a number no platform reports: what the next subscriber costs, channel by channel, at today's spend. Ad platforms report the average cost per acquisition, and the average says nothing about the margin. That gap is what marketing mix modeling fills.

Say your average cost per new subscriber across all channels sits well below what a subscriber is worth. The obvious conclusion is to spend more. It may be right, but the average cannot tell you. Every channel saturates: the first part of the budget reaches the people most ready to sign up, and each further step reaches people who are harder to convince. The average blends the cheap early subscribers with the expensive last ones.

So a channel can show a comfortable average while its last tenth of spend is already losing money. The reverse happens too: a channel with an unimpressive average can still have room to grow if the spend so far has been small. Saturation curves estimated per channel separate the two cases, because they show the marginal cost at every spend level.

The cheapest subscribers are not always the best ones. Some channels and offers bring in people who take a discounted first box and cancel before the second. Others bring fewer sign-ups who stay for years. A budget set on cost per sign-up keeps moving money toward the first kind.

There are three ways to handle this in a marketing mix model, in rising order of effort:

• Model on value instead of count. Use the revenue or gross margin from new subscribers as the outcome, rather than the number of sign-ups. Channels are then credited with the value they bring, and a channel full of one-box subscribers stops looking cheap.
• Model a later step. Use subscribers who reached their second or third delivery, or who stayed past a trial, as the outcome. That removes the people who were never going to stay. The delay between marketing and outcome gets longer, and the model takes that delay into account as part of its setup.
• Run two models side by side. One on sign-ups for reading the week, one on value for budget decisions. When the two point in different directions for a channel, that difference is the information you need.

Which one fits depends on your data. A value-based outcome needs a sales feed where each new subscriber's revenue can be tied back to the week they joined. Where that link exists, it is usually worth using. For many businesses the channels that bring revenue are also the ones that bring profit, and one model is enough. Where the mix differs, the second model earns its place.

Most of the work in a subscription model goes into deciding what the model should explain. These are the choices that matter most.

One more question comes up often: SMS and email to former customers. SMS has a cost per send, comparable with a media cost, so with enough variation it goes in as a paid channel. Regular newsletters to the whole base have no media cost and little variation, so they belong to the baseline.

Subscription demand has a rhythm. In many categories the start of the year and the weeks after summer bring a wave of people looking to start something new, and other periods are slower. In high season, marketing converts more easily, so the same spend wins more subscribers.

That leads to a result that surprises teams used to flat budgets: the months where you spend the most can also be the months with the lowest cost per subscriber, because demand is doing part of the work. A model that separates seasonal demand from the channel effect shows in which months the next part of the budget works hardest, and the budget can follow that instead of being the same every month. See how channel response can change through the year.

Subscription and DTC brands often put a large share of the budget into influencers, podcasts and video, where few people click a tracked link. Discount codes catch a fraction; many people see the content and sign up days later through search or by typing the address. Measured on clicks and codes, these channels look weak and get cut.

A marketing mix model sees them differently. It relates the spend and timing of each influencer push or podcast run to the movement in total new subscribers, whether or not anyone used the code. When a push lifts search sign-ups the week after, the model credits that lift to the push. What it needs is spend per period, so influencer and podcast costs go into the same structure as the digital channels. Read how to measure influencer marketing with MMM.

Odins is a managed marketing mix modeling platform. We connect the ad platforms and your subscription data, build and maintain the model, and review every recommendation before it reaches you. Data flows in continuously and the model is retrained once a month on the latest data, while scenarios and forecasts are yours to run whenever you need them, in the platform or from your own AI assistant.

For a subscription business, that means the budget level is set from your own ceiling for what the next subscriber may cost, channels are compared on marginal cost rather than averages, and the questions in this guide are answered on your own data. Moving budget toward channels with headroom typically gives 5 to 15% more effect from the budget you already have. Companies like NutraQ and Kokkeløren run their marketing budgets on Odins. See the managed service.