Cost per approved order: the number to scale against
A simple unit-economics model that turns payout, approve rate and lead cost into a break-even number you can hand to a media buyer.
Unit economics
Know the break-even line
Decision summary
- Break-even CPL = payout × approve rate. Everything above it burns money.
- Payouts in the catalog run $14–$34 per confirmed lead depending on the vertical and volume.
- Judge a test on approved orders, not clicks, and never on fewer than a few dozen leads.
- Scale in steps of roughly 20–30% of daily budget so the source keeps its learning intact.
The model in four lines
Every COD campaign reduces to the same arithmetic. Write it into a sheet once and every later decision becomes mechanical.
- Approved orders = leads × approve rate
- Revenue = approved orders × payout
- Cost per approved order = spend ÷ approved orders
- ROI = (revenue − spend) ÷ spend
Worked example
Take a $24 payout and a 47.8% approve rate on $1,200 of spend at a $4.20 CPL. That is 286 leads, roughly 137 approved orders, $3,288 in revenue and $2,088 of profit — about 174% ROI, with a cost per approved order near $8.80 against a $24 payout.
Now hold everything constant and let the CPL drift to $9.80. The same $1,200 buys 122 leads, about 58 approved orders and $1,392 of revenue. The margin has collapsed even though the campaign still 'converts'.
| Input | Healthy | Drifting |
|---|---|---|
| Spend | $1,200 | $1,200 |
| CPL | $4.20 | $9.80 |
| Leads | 286 | 122 |
| Approve rate | 47.8% | 47.8% |
| Approved orders | 137 | 58 |
| Payout | $24 | $24 |
| Revenue | $3,288 | $1,392 |
| Cost per approved order | $8.80 | $20.60 |
| ROI | 174% | 16% |
Break-even CPL, the number to pin above your desk
Break-even CPL is payout multiplied by approve rate. At $24 and 47.8% it is $11.47. At $16 and 40% it is $6.40. Anything above that line loses money on every single lead, and no amount of volume repairs it.
Set your working target at roughly 55–65% of break-even so there is room for approve-rate variance, refused parcels on higher-ticket products and the days when the auction is simply expensive.
Testing cadence that does not lie to you
Most 'winning' creatives are noise. A test needs enough approved orders to be real, and approved orders arrive a day or two behind the clicks because the call-center works on its own clock.
- Run three to five angles per test, one variable at a time.
- Wait for the confirmation lag before judging — day-one CPL is a leading indicator, not a verdict.
- Kill on cost per approved order, not on CTR or CPL.
- Keep one control creative running so you can tell a market shift from a creative decay.
Scaling without resetting your learning
Once a campaign is profitable on approved orders for several consecutive days, raise the daily budget by about 20–30% at a time and let it stabilise before the next step. Duplicate into a new audience rather than stacking budget on a single fragile ad set when the source starts to fatigue.
Tell your manager before a big step up. Caps, call-center capacity and stock are real constraints on COD, and a spike that outruns fulfilment shows up in your approve rate a day later.
FAQ
Common questions
Next decision
Apply the method to a live offer
Browse the current catalog, then confirm the exact payout, cap, traffic rules and operational capacity with your manager before scaling.