Subscription forecasting:
Campaigns when demand arrives
Subscription demand has an annual rhythm: resolution January, summer slump, gift-subscription December - plus news cycles that overlay everything. A forecast model makes this rhythm plannable and places budget and offers into the windows where they work.
01 The problem - The budget follows the calendar, not demand
Evenly spread campaigns buy subscriptions when they are expensive
In many houses subscription marketing runs in fixed waves: quarterly campaign, autumn push, year-end sprint. But demand does not follow the campaign plan - it follows seasons, holidays and news cycles. Piano observes the strongest concentration of promotional sign-ups between November and January; in summer the same campaign runs into a void.
Your own subscription history contains this rhythm in full: new subscriptions, cancellations, trial starts per week over years. A time-series model decomposes it into seasonal curve, trend and event effects - and delivers a reliable demand forecast per week with an uncertainty band.
Knowing your own seasonal curve means buying subscriptions in the demand peak instead of against the summer slump.
02 The model - Your house's own demand curve
Seasonal time-series models on your own subscription history, enriched with school-holiday, public-holiday and event features
Stärkste Nachfragefenster der nächsten 12 Monate: KW 02-04 (Januar), KW 46-51 (November/Dezember), KW 37 (Schulstart) MAPE (Backtest, letzte 12 Monate): 8.4%
The annual rhythm differs by house: a regional title with strong local sport has different peaks than a business outlet. That is why the model learns the curve from your history instead of industry averages - and at the same time flags which spikes were event effects (an election, a major event) and will not repeat on their own.
03 Business impact - The same budget, more sales
The lever: 15 percent more impact from the acquisition budget through timing and targeting
No round number: every assumption comes from the sample publisher and is stored centrally. With your real figures only the input changes, not the method.
| Item | Value |
|---|---|
| Subscription marketing budget / year | €600,000 |
| Baseline: Budget baseline | €600,000 |
| Lever: Impact gain through timing + targeting (15%) | 15 % |
| Result: budget effect / year | €90,000 |
Assumptions of a sample publisher - in a real project your data replaces these values.
Campaigns in the demand peak lower the cost per sale - and the subscribers won there come of their own accord, not through discount depth. That pays straight into durability: Mather Economics shows that short, moderate introductory offers with a clean step-up to full price maximise lifetime value - exactly this steering becomes plannable with a forecast.
04 Next steps in your publishing house
From your history to a rolling campaign plan
New subscriptions, cancellations and promotions of the last 3 to 5 years from the subscription system - the model needs nothing more.
The model back-casts the last 12 months. You see the hit rate before you base a decision on it.
Rolling 12-month forecast with marked demand windows - as the basis for budget, offer and staffing plans in circulation sales.
What this means for your media house is covered by our AI consulting for publishers.