How to Kill a Losing Promotion Before It Launches: A Pre-Launch Test for UK Mid-Market CPG
The cheapest promotion is the one you never run. Most trade promotions fail to break even, and most mid-market CPG brands only find out which ones failed weeks after the retailer has taken the deductions. A pre-launch test flips that order. It checks five things before the promotion is committed: legality, true incremental volume, fully loaded cost, break-even uplift, and history. That's how a £25M UK CPG brand saved £180K by cancelling two promotions before they launched (case studies).
This is the practical follow-on to our piece on AI pricing and promotion optimisation for mid-market UK CPG. That post argued you don't need an enterprise RGM platform. This one shows what to do instead, on Monday morning, with the data you already have.
Why does pre-launch evaluation matter more in the UK right now?
Because the promotion toolkit has shrunk and the tolerance for mistakes has shrunk with it.
The restriction of HFSS products by volume price came into force on 1 October 2025, covering multibuys and "free" mechanics on in-scope less healthy food and drink. It applies to all businesses selling food and drink into England, whether or not they're registered there, with businesses under 50 employees exempt. The next milestone lands this month: retailers have until 30 September 2026 to sell through existing HFSS products with promotions printed on the pack. Wales is stricter. Its restrictions came into force on 26 March 2026, and there retailers can sell on-pack promoted products but must not apply the promotion at the till, with no transition period. HFSS promotion, placement and advertising: March 2026 update +3
The rules may still move. The government has said it expects to repeal the multibuy restrictions, but has also called the Nutrient Profiling Model "plainly out of date" and signalled an update likely to bring products currently outside HFSS into scope. For planning, that means two things: the multibuy lever may come back, and more of your range may fall under the rules when it does. Lexology
Meanwhile, shopper behaviour has shifted. In Vypr polling reported at the start of the ban, 39% of shoppers said they'd buy fewer HFSS products once volume promotions disappeared, and only 27% were still happy to buy HFSS items off-promo, down from 61% in 2022. When fewer mechanics are available and each one carries more weight, every promotion needs to earn its place before it runs. BakeryAndSnacks.com
What does the five-question pre-launch test look like?
Run every promotion through these five questions before it's committed to the retailer. If any answer is "we don't know," that's the finding.
1. Is the mechanic still legal for this product, in this nation?
Check whether the SKU is in an HFSS category, whether its nutrient score puts it in scope, and whether the mechanic is a volume promotion. Straight discounts such as "50% off" or "save £1", free samples and meal deals are outside the volume price restrictions. On the manufacturer side, brands have been unable to print new packaging carrying non-compliant promotions since 1 October 2025. Pack artwork is now a compliance decision, not just a design one. Mills & ReeveHooleybrown
2. What's the true incremental volume?
Headline uplift isn't incremental volume. Strip out forward-buying by the retailer, pantry-loading by shoppers (the post-promotion dip), and cannibalisation of your own range. If your baseline forecast is weak, this question fails first. Our breakdown of where AI demand forecasting fails in UK food and drink covers why promotions are one of the four places baselines break.
3. What's the fully loaded cost?
Funding per unit is the smallest part of the picture. Add feature and display fees, extra logistics and cost-to-serve, and the deductions that arrive later. Execution drift is real. In one worked industry example, a BOGO promotion estimated at $100,000 ended up with $140,000 deducted because the retailer kept the discount running a week past the agreed end date. Cpgvision
4. What uplift does it need just to break even?
This is the question most teams skip, and it's the one that kills most losing promotions. See the worked example below.
5. What happened last time on this retailer, mechanic and SKU?
The best predictor of a promotion's outcome is its own history. For most mid-market brands, that history is spread across EPOS data, retailer portals, shipment records and a deductions spreadsheet that nobody has joined up.
How much uplift does a promotion need to break even?
More than most commercial teams assume. Here's an illustrative calculation with round numbers. It isn't drawn from any client.
| Input | Value |
|---|---|
| Baseline volume over a two-week promotion | 20,000 units |
| Manufacturer margin per unit at normal price | £0.60 |
| Promotional funding per unit | £0.40 |
| Margin per unit during the promotion | £0.20 |
| Fixed feature fee | £5,000 |
Without the promotion, those 20,000 units earn £12,000. During the promotion, every unit earns only £0.20, and you've also paid £5,000 up front. To match the no-promotion profit, you need (£12,000 + £5,000) ÷ £0.20 = 85,000 units.
That's 4.25 times baseline, a 325% uplift, just to break even. And that's before subtracting the post-promotion dip or any invalid deductions. This arithmetic is why the <a href="https://www.strategyand.pwc.com/gx/en/insights/2017/zero-based-trade-for-cpg-leaders/zero-based-trade-for-cpg-leaders.pdf">industry benchmark</a> finds most trade promotions don't pay back. Deep funding on high-baseline SKUs is structurally hard to recover.
The test is simple once written down: if the historical uplift for this retailer, mechanic and SKU is well below the break-even multiple, don't run it. Redesign it or put the money somewhere else.
Where does AI actually help, if not through a platform?
AI earns its place on questions 2 and 5, the ones that depend on joining messy data.
For a mid-market brand, the constraint is rarely the model. It's that promotion history lives in five places that don't talk to each other. AI helps in three specific ways:
- Reconciling history automatically. Matching past promotions to EPOS uplift, shipments and the deductions that followed, at SKU × retailer × mechanic level.
- Building a credible baseline. Separating true incremental volume from forward-buying and pantry-loading, so question 2 has a real answer.
- Flagging risk at planning time. Surfacing the promotions whose historical uplift falls short of their break-even multiple, before the joint business plan is signed.
Deductions are the other half of the same data. Retailer deductions can drain up to 3% of a CPG brand's revenue, wiping out 10 to 30 percent of profit margins, according to one vendor's figures. It's worth treating that as a range, since no public industry-wide benchmark exists for deduction load or invalid share. Either way, reconciling deductions against promotions builds the history that makes pre-launch evaluation possible. Our guide to AI deduction recovery for UK FMCG covers the mechanics. It's also why the work can start narrow: a £40M UK food brand recovered 60% of previously unchallenged deductions in eight weeks without building the data warehouse it had been told was a prerequisite. Glimpse Secures $10M to Automate Deduction Management for CPG Brands - Food Industry Executive +2
How do you start without a seven-figure platform?
Start with a diagnostic, not a system. The AI FlightCheck™ diagnostic is a fixed-price, 2–4 week assessment (£9,000) that tells you which promotions in your calendar are unlikely to break even, and ends in a 90-day action plan. Where it finds a build worth doing, such as an automated pre-launch evaluation tied to your planning cycle, that scopes into the AI FlightPath™ Sprint, which ships production AI inside ten weeks. AI Navi
If you need to win the budget first, our guide to presenting AI ROI to your board covers how to build the case without vendor projections. For the wider picture across all five RGM levers, see AI in revenue growth management for FMCG.
FAQ
How do you know if a trade promotion will be profitable before it runs?
Calculate the break-even uplift (baseline profit plus fixed costs, divided by margin per unit during the promotion), then compare it with the historical uplift for the same retailer, mechanic and SKU. If history falls well short, the promotion is likely to lose money.
What uplift does a promotion need to break even?
It depends on funding depth and fixed fees, and it's usually far higher than teams expect. In the illustrative example above, a promotion cutting per-unit margin from £0.60 to £0.20 with a £5,000 fee needs 4.25 times baseline volume just to break even.
Does the HFSS multibuy ban apply to manufacturers or retailers?
The restrictions apply to businesses selling to consumers in England with 50 or more employees, so retailers carry the compliance duty at the point of sale. Manufacturers are affected through pack artwork, promotion design and joint business plans, since in-scope volume mechanics can't run through qualifying retailers.
Can a mid-market brand predict promotion ROI without an enterprise RGM platform?
Yes, if the promotion history, EPOS data and deductions can be joined at SKU × retailer × mechanic level. That reconciliation is the hard part, and it's where a scoped AI build pays back faster than a platform licence.
Are straight price discounts affected by the HFSS rules?
No. Discounts such as "50% off" or "save £1" fall outside the volume price restrictions, which is one reason many brands have shifted mechanics since October 2025.