Picture this: Monday trade review and the dashboards look great. Last week’s promo “over delivered” with more orders, more redemptions, more engagement. Then finance drops the margin view into the deck and the room goes quiet; all that activity barely moved profit, and in some segments it actually went backwards. You did not buy growth; you just made the same customers cheaper.
According to recent trade promotion analyses, more than 60% of promotions actually destroy value rather than create it.
If that feels familiar, you are not on your own. Most brands are running promotions and loyalty incentives under pressure from rising acquisition costs, softer demand, new competitors and constant need for quick wins, so you reach for the levers that move fast: money off, points boosts, richer rewards.
Research on discounting shows that frequent, deep discounts can cut overall profit by 3–10% globally, even when headline sales look strong.
The hard part is telling which deals genuinely change behaviour and which just keep everyone busy while margin quietly drains away.
Underneath, you can probably see the pattern, but it is buried in noisy reports and celebratory recap decks. Some offers are clearly training people to wait for a code, some loyalty members behave like professional discount hunters, yet success is still framed in redemptions and top line lift.
Analysts looking at promo databases in CPG found that around one third of total volume, and almost two thirds of promoted sales in a typical event, can be subsidised sales to people who would have bought anyway.
The truth is, the data you already have is enough to separate genuine incremental lift from pure subsidy, the sales you are paying for even though they would have happened anyway, so you can start switching off low ROI deals, dial back over incentivised segments, and put budget behind behaviour shifts that actually grow the business.
Too many brands are paying people to shop, dine, or subscribe when those people were going to do it anyway. If that’s you – you’re ignoring
what your data is already telling you.
Let’s break this down a little.
Subsidy vs real lift: plain English
Think of every promotion as having two outcomes:
- Subsidised sales: people who take your offer but would have bought anyway at full price or with a lighter incentive.
- Incremental lift: extra visits, orders, or upgrades that happen only because the offer existed.
In value terms:
- Subsidy is pure margin burn; it changes who pays what, not whether they show up.
- Incremental lift is the only part of promo performance that should justify budget, because it is the part that would disappear if you switched the promo off.
When your subsidised share is high, the data is effectively saying: you are paying people to come anyway.
Benchmark work in CPG shows that, on average, around 35% of all sales on promotion are subsidised rather than incremental.
Core metrics to stop paying anyway
You do not need a PhD or a marketing mix modeling (MMM) project to see subsidy; a handful of simple metrics will get you most of the way.
Baseline vs promo performance
- Baseline: what this cohort or channel usually does without a deal (orders per user, AOV, visits, conversion).
- Promo period: the same metrics when the offer is live, ideally vs a small holdout or control that does not see the offer.
Incremental lift
- At its simplest: Incremental lift=Promo performance−Baseline
- Or, when using control tests: Lift=Treatment−Control on conversion, visit or usage rates.
Subsidy ratio
- If 40% of your orders would have happened anyway and all you did was cut price on them, that is roughly 40% of promo volume subsidised.
- CPG practitioners report typical ranges of 15–35% of total volume subsidised and up to 60% or more of promoted volume subsidised for some events, so this is not a corner case.
Discount cost vs profit impact
- A 20% discount rarely costs you 20%; on a 40–50% margin product, multiple pricing case studies show it can wipe out around one third to one half of your profit on that transaction.
- One worked example shows a 10% price cut on a 30% margin product producing a 33% drop in profit, which then requires roughly 50% more volume just to break even.
These basic metrics, sliced by segment and channel, are enough to highlight where you are over paying for very little incremental change in behaviour.
Simple analyses that expose low ROI deals
Here are straightforward analyses you can run in a spreadsheet or BI tool that quickly reveal where you are buying outcomes you would get anyway.
1. Baseline vs offer performance by segment
Objective: see which segments move meaningfully when you dangle money, and which barely budge.
- Step 1: Pick a past promo (for example “20% off first order” or “£5 off next visit”) with clear start and end dates.
- Step 2: Define segments such as new vs existing, high vs low frequency, high vs low basket, app vs web, high vs low intent (for example already in checkout).
- Step 3: For each segment, compare:
- Orders or conversions per 1,000 customers in a pre promo window (baseline).
- The same metric in promo window, adjusting for seasonality if needed.
Where to look:
- Segments where promo conversion is almost identical to baseline: you are mainly subsidising.
- Segments where promo conversion is materially higher than baseline and the extra revenue outweighs discount cost: these are your genuinely responsive, high ROI targets.
2. Intent based “would have bought anyway” check
Objective: find customers you are discounting at the exact point they were already committed.
Step 1: Tag sessions by intent:
- High intent: users who reached checkout, pricing or basket pages; opened “your order is ready” or “renewal reminder” flows.
- Medium or low intent: casual browsing, early funnel, non critical app usage.
Step 2: Compare conversion and average discount for high intent vs low intent cohorts.
Signals you are paying people to come anyway:
- High intent users show strong conversion even without a discount, and their lift when you add a deal is small.
- A large share of discount cost is concentrated on high intent events (for example adding a code at checkout) rather than nudging low intent users earlier in the journey.
The fix is simple: reserve deep incentives for low intent or low value cohorts; use lighter or no incentives where the behavioural data already screams they are coming.
A real ecommerce test shared publicly showed higher discounts actually delivered worse conversion than control at 25–35%, which is a classic sign of over use.
3. Net revenue per visitor (not just conversion rate)
Objective: stop chasing conversion lift that actually destroys value.
For a promo test (A vs B or promo vs no promo), calculate:
- Revenue per visitor or per recipient.
- Discount cost per visitor (average discount times conversion rate).
- Net revenue per visitor = revenue per visitor minus discount cost per visitor.
Patterns to watch:
- Offers that spike conversion but reduce net revenue per visitor are pure ROI traps.
- In one SaaS and ecommerce guidance piece, teams found that campaigns with the highest apparent conversion lift were not the ones that maximised profit once discount cost was included, underlining the need to move beyond surface metrics.
If your dashboards report conversion rate and incremental revenue without net revenue per visitor or profit, they are blind to subsidy.
4. Cohort retention of discounted vs full price customers
Objective: see whether discounted customers are worth less over time.
Build cohorts such as “acquired with 30%+ discount” and “acquired with low or no discount”, then track:
- Retention or repeat purchase rate over 6–12 months.
- Expansion revenue or average order value over time.
What the market data shows:
- Multiple loyalty and pricing case studies point out that heavy discount seekers tend to be less loyal, more price sensitive, and quicker to churn than full price cohorts.
- Across subscription and membership businesses, aggressive upfront deals are correlated with lower lifetime value and weaker upsell, a classic case of buying the wrong customers too cheaply.
If your best performing acquisition campaigns are built on heavy, always on discounts, this cohort view often reveals that you are importing churn risk and training people to only show up when paid.
Spotting over incentivised segments
Once you have the basic views, specific segments tend to pop out as over incentivised.
Here are some usual suspects, and the data signals that give them away.
| Segment type | What the data looks like | Why it screams “over incentivised” |
| Loyal high frequency users | High visit or usage baseline, minimal incremental orders during promos, heavy discount usage | You are repeatedly cutting margin on behaviour they already show, which aligns with findings that a large share of promo volume is subsidy not lift. |
| High intent sessions (checkout or pricing) | Strong natural conversion, tiny lift with vouchers added at end of funnel | Discounts are applied too late and mostly subsidise committed buyers, just as ecommerce tests show deep discounts failing to outperform no discount at all. |
| Deep discount acquisition cohorts | Huge sign up spikes when hefty offers run, then weak retention and low expansion | You attract deal seekers who disappear or stay low value, a pattern flagged in loyalty and LTV studies. |
| VIP segments with blanket perks | High margin cost per head, small behavioural difference vs non VIP peers | You are confusing recognition with financial incentives; industry commentary warns that discount heavy loyalty programs often dilute value more than they drive it. |
None of these require complex modelling; most can be surfaced with:
- A few cohort tables (by segment, intent and acquisition offer).
- Simple A/B or holdout tests for key promos.
If a segment does not budge in behaviour when you take the money away, you have just identified subsidy.
Turning analysis into guardrails
Finding subsidy is only half the job; you need to bake the insight into the way you design and approve offers.
Practical guardrails that data teams and marketers can agree on:
- Define an acceptable discount cost ratio (discount as percentage of revenue) by channel or lifecycle stage, and block offers that exceed it without a clear incremental case. Pricing and revenue management guides consistently recommend setting these bands explicitly instead of relying on gut feel.
- Require every major promo to have a control or holdout, with decisions based on incremental lift and net revenue per visitor, not just topline sales. Incrementality playbooks across retail and digital products point to control groups as the single most important ingredient in avoiding subsidy.
- Build simple “do not over incentivise” rules, for example no discounts on already discounted items, no vouchers injected at checkout for users with high loyalty scores, and a cap on frequency of heavy deals per user. Analysts warn that predictable discount cycles train customers to wait, which is exactly what these rules are designed to break.
- Treat intent and loyalty as levers: richer incentives for low intent, at risk or high potential segments; lighter recognition based benefits (access, experiences, soft perks) for your already loyal base. Loyalty experts increasingly argue that programs built mostly on discounts underperform those that balance financial rewards with status and experience.
I’m not suggesting to stop discounting; but do think about paying for behaviour you would get for free and re allocate that budget to the places where incentives genuinely change what people do. In a world where more than half of promotions risk destroying value, using your own data to separate lift from subsidy is one of the fastest ways to protect margin and still grow.
If you know something is off but cannot prove which promos are just burning margin, that is exactly where Massive Rocket comes in. We connect your data and engagement stack so you can see true incremental lift, redesign offer rules, and stop over paying the customers who were coming anyway.
If you are asking “Which deals should we kill, and which should we double down on?”, we can help you answer that with numbers, not gut feel. Get in touch with Massive Rocket and let’s turn your offer budget into a growth engine instead of a subsidy fund.
