Ordering for a number you will not know for weeks
The brand needs the order for the May box by the end of March. Right now you have a subscriber count in the app, but it includes people who will skip, people whose cards will fail, people who will cancel after the April box, and it misses the signups that will come in during April. So you take the active count, add a bit for new signups because a promotion is planned, take a bit off for churn, and round up to the brand's case size.
In May you either have a pallet of spare body wash or a scramble to find thirty more units when the promotion did better than expected.
The active count is the wrong starting number
The number in your subscription app is a snapshot of today, but the box you are ordering for ships in six or eight weeks. Between now and then, several predictable things happen, and each of them is visible in your own history.
- A share of active subscribers skip in any given month, and that share changes with the season.
- Some renewals fail and are never recovered.
- Some subscribers cancel after each box, more after certain boxes.
- New signups arrive, with bumps when you run promotions or when gift season starts.
- Prepaid subscribers are certain; monthly ones are not.
None of this needs guesswork. It needs someone to turn past months into a rate for each of those movements.
What ordering by feel costs
Over-ordering ties money up in stock that has to be stored, sold off in a clearance or squeezed into a later box where it may not fit the theme. Under-ordering means buying top-ups at a worse price, splitting the box into versions, or delaying dispatch. Neither is catastrophic once, but together they erode the margin on every box.
It is also stressful. The owner carries the forecast in their head every month, and nobody else can check it.
The forecast we build from your own history
- We pull your past months from the subscription app: active count at each point, skips, pauses, failed renewals, recoveries, cancellations and new signups by source.
- Rates are worked out for each movement, by month of the year and by subscriber type, such as monthly, prepaid and gift.
- For any future box, the forecast starts from today's subscribers, applies those rates and adds expected signups from your plan, including promotions you enter.
- It shows a likely number and a range, not a single figure pretending to be certain.
- It updates every day as real skips, signups and cancellations come in, so a quantity you ordered can be compared with the latest view.
- Variant counts, such as sizes and exclusions, are forecast from the same subscribers, so each variant has its own number.
| Movement | Where it comes from | How it is used |
|---|---|---|
| Skips and pauses | Subscription app history | Rate by month of the year |
| Failed and recovered renewals | Payment provider and app | Net loss rate per renewal |
| Cancellations | Subscription app | Rate after each box type |
| New signups | Shop orders, marketing plan | Baseline plus planned promotions |
| Prepaid boxes | Subscription app | Counted as certain until the plan ends |
The forecast gives you a number and the reasoning behind it. How much margin to add, and how you round to case sizes, is your call.
Ordering with the forecast open
When the brand asks for the May order, the forecast shows a likely count and a range. You can see why: the April skip rate is usually higher because of Easter, prepaid subscribers make up a fixed block, and the planned promotion adds signups based on how your last promotion performed. You order within the range, rounded to case size, and write the reasoning down in the order note.
As April progresses, the forecast tightens. If it drifts outside the range, you know early enough to ask for a top-up.
Signs you are ordering by feel
- You order from the current active count plus a margin you cannot explain.
- Leftover stock builds up after most boxes.
- You have bought emergency top-ups at a worse price.
- Variant quantities are ordered as a guess.
- Only one person knows how the order numbers are worked out.