Why a loyalty programme changes the calculator's inputs
On the default page, spend is a media budget and every sale counted against it is one the campaign is credited with. A loyalty scheme has no media budget. Its cost is made of rewards, the platform that tracks them, the people who run it and the messages that keep members informed, and the reward line grows with every purchase members make, whether the scheme caused that purchase or not.
That is the problem the guide to loyalty programmes puts at the centre of its economics. A programme pays when the extra margin from behaviour it changed is larger than the rewards actually claimed plus the cost of running it, and that running cost includes the rewards paid to regulars for visits that would have happened without the scheme, the cost business cases most often leave out. The same guide points to the break-even calculator to frame the question of how much extra behaviour the programme must create.
Two habits from the default page would answer it wrongly. Counting every member order as a programme sale credits the scheme with the habits of customers who were already loyal, and the guide warns that schemes mostly attract existing heavy buyers. And valuing only the rewards claimed so far, or assuming a share will never be claimed, makes the cost look smaller than the debt the business has taken on.
The inputs for a loyalty programme
| Calculator input | On the default page | For a loyalty programme | Where to find it |
|---|---|---|---|
| Average order value | Average order value of the campaign's sales | Average value of members' orders | The order system, members only, over the pilot window |
| Gross margin | Everyday gross margin | Contribution margin on members' orders before any reward | Finance margin for the products members buy |
| Campaign spend | Media spend for the campaign | The programme's cost for the period: rewards redeemed, rewards earned and still owed valued in full, technology, staff and communication | The reward ledger, the platform invoice and staff time |
The spend row is the derivation that matters. Start from the reward ledger for the year: every reward claimed, plus every reward earned and not yet claimed, each valued in full, with no reduction for BreakageThe share of issued loyalty rewards that customers never redeem, because points expire, balances are too small to use or members forget.. The loyalty programme pilot test values outstanding rewards in exactly this way. Add the platform or app, the staff time spent explaining and redeeming, and the printing and messages. Because the reward line counts rewards earned on habitual purchases too, the reward on purchases regulars would have made anyway is already inside spend; nothing needs to be added for it separately.
Order value and margin come from members' own orders, because they are the orders the rewards attach to. The margin is Contribution MarginRevenue left after every cost that varies with each sale. before any reward, since the rewards are already counted in spend; taking them off the margin as well would charge the programme twice.
Entering the loyalty programme inputs step by step
Fix the period and the eligible customers
Choose a year, or the pilot's reading period, and count the customers eligible for the scheme in it. That count turns the calculator's answer into a figure per customer later.
Build spend from the ledger and the invoices
Total the rewards claimed and still owed at full value, then add the platform, staff and communication costs for the same period.
Take order value and margin from members' orders
Use members' average basket from the order system and the margin finance reports for the products they buy, before any reward is deducted.
Enter the inputs and translate sales needed
Enter the three figures. Note the sales needed as the number of extra orders, spread it across the eligible customers, and keep that rise per customer as the gap the comparison group must show.
A worked loyalty programme with illustrative inputs
Tulmuth Bakehouse costs a stamp card for its six bakeries
Tulmuth Bakehouse, an invented group of six bakeries, plans a stamp card kept in its ordering app. Members' baskets average £9 at a contribution margin of 60%, so each order earns £5.40. For a year, rewards earned, whether claimed or still owed and valued in full with nothing assumed for breakage, come to £9,600; the app costs £3,000; staff time and printing add £1,800. Spend is therefore £14,400.
The calculator returns a break-even CPA of £5.40, a break-even ROAS of 1.67 and 2,667 sales needed: 2,667 extra orders a year beyond what customers would have bought without the card. Tulmuth has 4,000 eligible regulars, so each needs to visit about two-thirds of a time more per year, roughly 0.67 extra orders each, or 0.33 over six months. Much of the £9,600 goes on stamps for visits regulars would have made anyway; that cost is already in spend.
Tulmuth pilots the card in three shops and holds out three matched shops. After six months, long enough for stamps to be earned and rewards claimed, regulars of the pilot shops have placed about 0.5 more orders each than regulars of the held-out shops, with a range of roughly 0.2 to 0.8. The central figure clears the 0.33 the half-year needs, but the lower end does not, so Tulmuth keeps reading for another quarter before offering the card in all six shops.
Campaign Break-even Calculator
Open the Campaign Break-even Calculator →Prefilled with the example's inputs: Average order value $9 · Gross margin 60% · Campaign spend $14,400 → Break-even CPA $5.40 · Break-even ROAS 1.67× · Sales needed to break even 2,667.
All figures belong to the invented bakery group. Lower spend in the embedded calculator to the rewards claimed so far to see how a case built on unclaimed rewards shrinks the extra orders it appears to need.
Interpreting sales needed as extra member orders
Start with sales needed. Every order in it has to be one that members place on top of the held-out group's buying, never a share of everything members order. The pilot test counts everyone assigned to the offered group, joined or not, and compares them with the held-out group, so the pilot's gap and the calculator's figure measure the same thing. A scheme whose members place many orders can still fall well short, because most of those orders were coming anyway.
Then divide by the eligible customers. The result is the rise in Purchase FrequencyAverage number of orders each customer placed in the period. the rewards must create for each customer offered the scheme, which is the first of the jobs the loyalty guide lists for a programme. Expressed that way, the target can be checked against what customers in the category already do: a rise that would ask regulars for far more visits than they plausibly make is a sign the design or the cost is wrong. Repeat Purchase RateShare of customers who placed more than one order in the period. in both groups shows whether the rise comes from more customers returning or from the same customers returning more often.
Read the cost behind the figure with the loyalty guide's warning in mind. The rewards line in spend includes margin given to members for visits they would have made anyway, which is the cost the loyalty guide says business cases most often leave out. When the extra orders needed look out of reach, the guide's answer is to move rewards towards behaviour the business would not otherwise get, not to trim the reward line by assuming members will not claim.
Finally, keep rewards owed at full value when the result is read. The diagnosis of loyalty schemes that leave retention unchanged treats a high rate of unclaimed rewards as evidence the scheme is failing, since rewards left unclaimed show members are not responding to them, and the loyalty guide calls a case that pays only through breakage one designed to disappoint members. Where customers reorder one product at regular intervals, the comparison between a loyalty scheme and subscribe-and-save may point to a standing order instead, whose economics the break-even page for subscription businesses reads over a payback window. The break-even page for direct mail applies a held-back group to a mailing in the same spirit as the pilot here.
Pitfalls when loyalty programme numbers go into the calculator
- Counting every member order as caused by the programme. Members were often the best customers before they joined; only the gap over a held-out groupA randomly chosen part of an audience kept from exposure to measure what exposure caused. belongs against sales needed.
- Assuming breakage to make the case pay. Lowering spend for rewards nobody expects to claim builds a case that works only if members feel short-changed.
- Leaving staff, technology and communication out of spend. A rewards-only cost understates the extra orders needed and hides the part of the bill that does not shrink when rewards are redesigned.
- Reading the pilot before rewards can be earned and redeemed. Early sales arrive before the reward cost does, so a reading taken too soon shows the benefit without the bill.
- Comparing members with non-members who chose not to join. Self-selected members flatter the scheme; compare everyone offered it with a similar group that was not.
