Most drive-to-retail campaigns launch without a number to hit. There is a budget, a rebate amount, a flight window, and a general hope that Meta will figure it out. Eight weeks later someone asks whether it worked, and the honest answer is that nobody set the bar beforehand.
The fix is to derive one target from your unit economics before a dollar is spent, then treat every metric after that as a diagnostic rather than a scorecard. Here is how that works end to end.
Start at the shelf. Take a four-pack in the chilled aisle at $12.99 with a $2.00 rebate offered. Subtract the rebate. Subtract operational cost and payout processing, call it $0.75. What remains is $10.24, which is the ceiling on what you can pay in media for one verified receipt.
Plan to $9.00 rather than the full $10.24. The buffer is deliberate. Positive margin comes later, once you know how the campaign behaves. The first job is designing economics that net to zero or better.
Then translate. Meta does not optimize on receipts. It optimizes on opt-ins, meaning the email captured on the offer page, so the receipt target has to become the metric the platform can actually chase:
$9.00 target per receipt × 12% expected redemption = $1.08 target opt-in CPA.
Check it. One hundred opt-ins at $1.08 is $108 of spend. Twelve redeem. Cost per receipt lands at exactly $9.00.
That $1.08 is the number the entire media operation exists to hit. Two adjustments make it yours rather than an example. First, $12.99 is the retail price and what your business captures is your wholesale margin on it. Reported gross margins across big food give you the range: General Mills at 33.6% for fiscal 2026, Campbell's at 27.3%, Hershey at 45.3%. Second, this is a one-purchase view. Breakage, meaning shoppers who bought after seeing the ad but never submitted a receipt, and repeat purchase rate both raise what the campaign is genuinely worth.
Hold the $9.00 receipt target constant and change only the redemption assumption. The affordable opt-in CPA swings completely:
A $1.08 opt-in CPA is meaningless at 10% redemption and a $3.60 one might be excellent at 40%. This is why the target gets set on the platform average and then corrected by real receipts within the first few weeks. It is the largest single source of error in the model and also the one the data resolves fastest.
Once the target exists, every other metric stops being a report card and starts answering one question: how close are we to target cost per receipt, and why.
Five layers, each with a benchmark and a job:
The job is not to optimize CPM or optimize CTR. It is to identify which layer is misaligned and apply the correct fix.
No single metric resolves on its own. Five patterns cover most of what you will see:
That fourth pattern is worth dwelling on. A stable top-funnel can mask a degrading bottom one, and the system will happily optimize toward easier conversions while basket quality declines and rebate-only receipts grow.
Watch behavior over time too, not just levels. Hook rate declining from week four is the earliest warning of creative fatigue, and CTR typically follows it down about two weeks later. Because new creative takes roughly two weeks to produce, waiting for CTR to break guarantees downtime.
Diagnosis tells you what is broken. Constraints tell you what is allowed. Only then does a lever make sense, and there are seven worth knowing: scale what is working, duplicate the winning pattern with new creators, refresh creative, adjust the offer, bias toward better audiences using receipt data, structure geographic tiers, and preserve accumulated learning.
Offer adjustment sits in that list rather than at the front of it. Reach for it when hook and hold are strong and CTR is weak, which says the value exchange is not landing. Even then, reframe the offer before increasing it, since a bigger rebate directly worsens the unit economics you just spent all that effort defining. Context for the judgment call: Ibotta found 62% of food shoppers require a discount of 25% or more to switch brands, so a $2.00 offer on a $12.99 pack is below the threshold that typically moves behavior.
However you pull a lever, five guardrails keep the read clean:
The floor is $3 of value back per media dollar. Healthy, the kind of program that renews, lands between $6 and $12. Those figures include breakage and repeat purchase, which is why they sit above the break-even design the target was built around.
For context on what the discipline is worth, McKinsey found CPG companies investing around 20% of revenue in trade promotions, with 59% losing money and 72% losing money in the United States, while best-in-class promotions returned five times more than the least efficient. Recent Circana work shows the same thing at brand level: one frozen entrée brand had roughly 60% of volume selling on promotion, of which about 40% was subsidized rather than incremental.
The gap between those outcomes is not better creative or better targeting. It is whether anyone defined the number before launch and then read the funnel against it.
Take your retail price, subtract the rebate and operational cost, decide what you can pay per receipt, multiply by expected redemption, and launch against that opt-in CPA. Then let the funnel tell you which layer is off, let the constraints tell you which levers are available, and change one thing at a time.
The full playbook, including the six pillars in the order a campaign actually lives them, the four system pathologies, and how receipt data feeds back into the weekly decision, is available here.
.svg){kind=icon}
.svg){kind=icon}
.svg){kind=icon}
.svg){kind=icon}
.svg)
.svg)
.svg)
