Why direct mail changes the calculator's inputs
The default calculator page imagines a paid campaign: a media budget, sales a platform attributes to it, and a margin that applies to each sale. Addressed mail has none of these. There are invoices for design, print, postage, handling and sometimes a rented list, and there are the business's own order records. It also brings a temptation the default page never meets: every recipient is named, so every mailed customer who buys can be looked up and credited to the mailing.
The guide to measuring direct mail calls that matching exercise matchback and treats it as an upper limit, not an effect. Mailing lists are chosen for people likely to buy, so many recipients would have ordered without the catalogue. What the mailing caused is the difference between the mailed customers and a random Holdout GroupA randomly chosen part of an audience kept from exposure to measure what exposure caused. drawn from the same list before the drop, and that difference is the only figure the calculator's sales-needed output should be compared with.
The same guide costs a mailing on margin rather than revenue, since extra sales on thin-margin lines can leave the mailing out of pocket, and that is the arithmetic the calculator performs once its inputs describe the mailing.
The inputs for a direct mail campaign
| Calculator input | On the default page | For direct mail | Where to find it |
|---|---|---|---|
| Average order value | Average value of the campaign's sales | Average value of the orders the mailed segment places by any route, net of tax, delivery and returns | The order system for the same segment over the last comparable response window |
| Gross margin | Everyday gross margin | Gross margin on those orders from the business's own accounts, as a share of order value, never revenue alone | Finance margin for the products the segment usually buys |
| Campaign spend | Media spend for the campaign | The segment's share of the mailing bill: design, print, postage, list rental and handling | Invoices from the designer, printer, mailing house and list owner |
The spend row changes most. Design is shared, so split it by the pieces each segment receives; print, postage and handling follow the pieces; list rental belongs only to segments drawn from a rented list. Order value counts orders placed by any route, because mailed customers rarely use only the channel the piece suggests.
One relationship is worth writing beside the table. Because spend is the number of pieces times the cost per piece, dividing the sales needed by the number mailed gives the cost per piece divided by the margin per order. That ratio is the break-even lift in order rate, and it does not depend on how large the segment is.
Where a mailing asks for an enquiry rather than an order, the break-even calculator page for lead generation applies, and a mailing selling a plan follows the break-even calculator page for subscription businesses.
Entering the mailing inputs step by step
Settle the segments and the held-back group before the drop
Apply the selection rules, remove everyone who may not be mailed, then draw a random share of each segment to hold back. The holdout test readiness checklist is a gate for this moment: the decision named, the groups flagged in the records and the reading date fixed. Run the calculator once per segment, because each will be judged on its own.
Build each segment's spend from the invoices
Add the segment's share of design, its print and postage for the pieces it receives, handling, and list rental if the names were rented. Leave out the held-back customers, who cost nothing to not mail.
Take order value and margin from the business's own records
Use the segment's own recent orders and the margin finance reports for what those customers buy. A margin taken from the whole range flatters a segment that buys mostly low-margin lines.
Enter the inputs and translate the outputs
Type the three figures into the calculator. Write the sales needed down as incremental orders, divide it by the number mailed to get the break-even lift in order rate, and record both in the campaign record before the first piece arrives.
A worked mailing with illustrative inputs
Druskelt Yarns sets the bar for its autumn catalogue
Druskelt Yarns, an invented mail-order seller of yarn and knitting supplies, mails its autumn catalogue to customers who have not ordered for over a year, in two segments, holding back a random 10% of each. Orders from these customers average £52 after tax, delivery and returns, at a gross margin of 45%, so each order earns £23.40. The catalogue costs £0.78 a piece to design, print, post and handle; no list is rented. The first segment, lapsed 13 to 24 months, has 7,200 mailed and 800 held back, so its spend is £5,616. The calculator shows a break-even CPA of £23.40, a break-even ROAS of 2.22 and 240 sales needed: 240 incremental orders, a lift of 3.33 points in order rate (£0.78 ÷ £23.40). The second segment, lapsed 25 to 48 months, has 6,300 mailed for £4,914 and needs 210 incremental orders, the same lift.
Ten weeks after the drop, the planned reading date, 15.0% of the first segment's mailed customers had ordered (1,080) against 8.0% of its held-back customers (64): about 504 incremental orders, with a range of roughly 360 to 650, comfortably above 240. In the second segment, 7.0% against 5.0% gives about 126 incremental orders, with a range of roughly 20 to 235, short of 210. Matchback would have credited the catalogue with all 1,521 mailed buyers. The first segment added about £6,180 of margin after the mailing's cost; the second lost about £1,970. Druskelt keeps the catalogue for the first segment and tests a £0.29 postcard on the second, which needs a lift of only about 1.24 points.
Campaign Break-even Calculator
Open the Campaign Break-even Calculator →Prefilled with the example's inputs: Average order value $52 · Gross margin 45% · Campaign spend $5,616 → Break-even CPA $23 · Break-even ROAS 2.22× · Sales needed to break even 240.
Druskelt and every figure in its mailing are invented for this illustration. Enter the second segment's bill as spend to see the sales needed fall in proportion while the lift needed stays the same.
Interpreting the break-even outputs for a mailing
Read sales needed first, and read it as incremental orders: the orders the mailed segment has to add over its held-back group, not the orders mailed customers place. The guide to measuring direct mail turns the holdout difference per person into an incremental total by multiplying it by the number mailed, which puts the result on the same scale as the calculator's figure.
Then turn that figure into the lift in order rate: sales needed over the number mailed, or the cost of a piece over the margin per order, is the gap the mailed group has to open over the held-back group. This is also how the direct mail format test weighs a catalogue against a cheaper piece: each piece's cost sets the margin it must produce, so a lighter piece can pay with a smaller lift.
Break-even CPA is the most the mailing can spend for each extra order. A cost per order from matchback, the bill divided by every mailed buyer, will look far below it and prove nothing, because matchback counts customers who would have ordered anyway. Break-even ROAS works the same way on revenue: incremental revenue over the mailing's cost has to reach it, and if mailed customers place larger orders as well as more of them, compare margin per person in each group rather than order counts alone.
Read the holdout difference with its range, on the date fixed before the drop, segment by segment. A segment whose range runs from below the bar to above it has not shown a result yet, as the guide to reading small numbers explains, and the incrementality testing guide covers how large the held-back group must be for the lift the calculator asks for to be visible at all.
Finally, the calculator cannot see value after the window. Lapsed customers brought back may keep buying, and the lifecycle marketing guide gives reactivation a stage of its own, measured separately; a later read of repeat orders in both groups shows whether that value is real.
Pitfalls when mailing numbers go into the calculator
- Entering coded or matchback orders as the result. Codes and tagged addresses are floors and matchback is a ceiling; only the held-back comparison says what the mailing changed.
- Leaving list rental, design or handling out of spend. A bill of print and postage alone sets the bar too low.
- One verdict for the whole list. The choice to be made is where the next mailing goes, segment by segment, so each needs its own run.
- Costing the mailing on revenue rather than margin. Revenue flatters offers on thin-margin lines; the Gross MarginShare of revenue left after the direct cost of the goods sold. input is what makes the calculator's answer a test of profit.
- Reading the result before the response window closes. Orders from a mailing arrive over weeks, and a figure read early and reported at its most flattering moment is noise.
