Promo calendar and promotion effectiveness
The promo calendar allocates promotions across weeks and the discount budget, and the post-promotion evaluation shows whether it earned extra profit or merely gave away margin. It decides which mechanics to repeat, which to drop and how much stock and money to allocate next period.
Why this document
A promotion lifts revenue but not necessarily profit: some customers would have bought without the discount, and some pulled forward demand from later weeks. That is why after a promotion actuals are compared with a sales baseline, not with last year, and the lost margin and advertising cost are counted. Marketing plans the calendar so promotions do not overlap or eat seasonal peaks. The finance director sees from the evaluation which mechanic pays back and cuts the budget of those that only give away discount.
At a glance
- What is this document and why is it needed?
- The promo calendar allocates promotions across weeks and the discount budget, and the post-promotion evaluation shows whether it earned extra profit or merely gave away margin. It decides which mechanics to repeat, which to drop and how much stock and money to allocate next period.
- Who owns the document and how often is it rebuilt?
- Brand: marketing and the commercial director, an annual calendar reviewed quarterly, agreed with the finance director. Retailer: category managers and marketing, a weekly calendar; an analyst reviews results within 2-4 weeks after the end.
- What mistakes are the most common?
- A promotion is judged by revenue growth → it looks successful while profit falls → judge incremental profit. Cannibalization and the post-promo dip are not counted → the result is overstated → include both corrections. A promo code goes to everyone → most buyers would have bought anyway → compute the self-selection share and the payback threshold.
For whom
- Brand
- A brand plans promo in three layers: its own channels (web, stores), wholesale (partner support, co-funding) and marketplaces (platform promotions). The main risk is training the shopper to wait for discounts and undermining the price at partners; hence price protection (a reference price) and calendar coordination with retailers.
- Retail
- A retailer plans promo by category and week with a limit on discount amount and weeks on promotion; it also runs bonus mechanics (points, cashback) and recognizes the points liability. Supplier and brand co-funding enters the calculation, not a “general budget”.
What it consists of
The calendar
- Week, mechanic (discount, multibuy, promo code, gift, points), category, depth, term, channel.
- Limits: discount amount in the budget, weeks on promotion per category, margin floors.
- Conflicts: overlapping mechanics (stacking), coincidences with platform sale days.
- The baseline of 40 units a week in the example is derived as follows: average sales over the last 8 weeks without promotions of 38 units × the seasonal index of the promotion window 1.05 = 39.9; for checking a control group of similar stores without the promotion is used (difference of differences), not just “last year”. A multibuy “two for 1.5”: the price per unit of RUB 3,750 at cost of 2,000 gives a margin of 1,750 instead of 3,000, so for payback sales must grow by 71.4% before cannibalization. A marketplace promotion: if the platform lowers the item price at the seller expense, it is allowed by contract with at least five business days notice to the partner and with its consent (289-FZ, Art. 9); add a “compensation to the platform” line to the promo calculation.
Post-event evaluation
- Baseline: sales without the promotion (earlier weeks, last year, a control group).
- Uplift, cannibalization (lost at neighboring items), pull-forward and the post-promo dip.
- Incremental profit after promotion costs and returns; elasticity from the actual.
Bonus mechanics
- Points accrual: the liability for unredeemed points; the breakage share is an estimate, not a fact.
- Promo code: the share “would have bought anyway” (self-selection) decides whether the discount pays back.
- Promotion results: actual increment = (Promo-period profit − Baseline profit) − Cannibalization − Pull-forward − Costs; in the example RUB 28,488, while a naive calculation without cannibalization and pull-forward would give RUB 88,680, an overstatement of RUB 60,192. Decisions after the results: repeat, change the mechanics (depth, duration, range) or stop; after the promotion watch the sales dip of the following weeks (pull-forward).
- Weekly promotion control: plan and actual sales against baseline, the share of promotional sales in revenue, stock of promoted items (so as not to run out in the second week), returns of promotional orders; a stop trigger: the second-week increment more than 30% below plan (an editable threshold).
Brand and retailer: what is computed differently
- Brand: brand promotions: participation in platform promotions (the price is often lowered by the platform at the seller expense), wholesale volume bonuses to the customer and compensation for price reductions on the partner shelf; all reduce revenue and enter the increment calculation.
- Retailer: retailer promotions: in-store mechanics (discount, multibuy, points, coupons), cannibalization of neighboring categories and pull-forward; stock for the promotion and the display are counted in advance so as not to run out in the second week.
Brand: how the document differs and what to add
- Brand: brand promotions are participation in platform sales (the platform often lowers the price at the seller’s expense), wholesale volume bonuses and compensation to the partner for shelf price reductions.
- Brand: a discount on the own site affects wholesale: a partner sees a price below its own and demands compensation; the calendar is agreed with partners in advance.
- Brand: when joining a marketplace promotion the net price after commission, logistics and the platform’s extra discount is computed; without it unit profit is overstated.
- Brand: a deep discount on the current collection reduces brand value: the permitted depth and frequency are written in the pricing policy.
- Brand: clearance of last collection remaining stock goes through the outlet, own site and closed sales: the channel is chosen by its effect on price in the main channels.
- Brand: promo codes and referral mechanics are assessed by the acquisition cost of a new customer and repeat purchases, not by the discount per order.
- Brand: the promotion budget is taken from marketing but affects returns and channel margin: the assessment is shown in the channel P&L, not only in the marketing budget.
Retailer: how the document differs and what to add
- Retailer: retailer promotions are category price reductions, card promotions, cross-sales and multi-buy mechanics; the effect is measured on ticket and traffic.
- Retailer: cannibalization within the network and shifting of purchases from adjacent weeks are assessed through control stores without the promotion (diff-in-diff).
- Retailer: an in-store promotion changes staff and merchandising: costs of extra shifts, printing and display go into promotion costs.
- Retailer: a loyalty-card promotion creates a points liability: the effect appears later in point redemption, so payback is computed with a lag.
- Retailer: joint promotions with a brand (cofund) share costs by contract: the brand share reduces the retailer’s promotion costs and enters the plan.
- Retailer: season end: planned price reductions by age and WOS follow a ladder (markdown ladder), not one deep discount.
How to read the result and which decisions follow
- Incremental profit is negative while revenue grew: the promotion bought turnover at the price of margin; it must not be repeated without changing depth or mechanics.
- Sales uplift is below that required for the discount depth: reduce depth or switch to a second-purchase mechanic; do not justify with a “brand awareness” effect without measuring repeat purchases.
- After the promotion sales fell below baseline in the following weeks: pull-forward is large; place the next promotion further from the previous one and compare over a period before and after.
Control checks
- The increment ties: period profit 327,560 − baseline 218,880 − cannibalization 27,360 − pull-forward 32,832 − costs 20,000 = RUB 28,488; required sales uplift 58.7% against actual 137.5%.
- The baseline is defined before the promotion by a method written in advance (past period, control stores); change the method and the result changes.
- Prices in the calculation exclude VAT, promotion returns are separate, promotion costs include the discount, advertising and handling.
- Typical traps: evaluation on revenue instead of profit, no cannibalization or pull-forward, the baseline taken from the best week, promotion returns set equal to normal ones.
Where the data comes from
| Data source | Fields |
|---|---|
| POS and marketplace | Sales by week and SKU before and after the promotion, price, discount, returns. |
| Budget and finance | Discount limit, promotion costs (ads, flyers, platform fee), co-funding. |
| CRM | Participants, repeat purchases, share of new customers, points. |
How to build it step by step
- Step 1. Define the goal: clear stock, acquire new customers, support a launch; the metric depends on it.
- Step 2. Compute the baseline: sales without the promotion over comparable weeks, adjusted for season and weather.
- Step 3. Set depth and term and check margin against the floor; compute the uplift needed for payback.
- Step 4. Enter it in the calendar; check mechanic conflicts and the discount budget limit.
- Retailer: Step 5. Fix a control group (stores or customers without the promotion) where possible.
- Step 6. After the promotion collect the actual: sales, returns, costs; estimate cannibalization and pull-forward.
- Step 7. Compute incremental profit and payback; assess the dip in the next 2-4 weeks.
- Step 8. Record the verdict (repeat, change, drop) and update the calendar rules.
Formulas
| Measure | Formulas | Note |
|---|---|---|
| Unit profit | u = (1 − Return share) × (Price − Cost) − Return share × Cost per return | Compute separately for the promotion (discounted price) and for normal time. |
| Incremental profit | Δ = Sold in promotion × u₁ − Baseline × u₀ − Cannibalization × u₀ − Pull-forward × u₀ − Costs | The “Promotion result: incremental profit” calculation. |
| Required uplift without costs | Uplift, % = (u₀ ÷ u₁ − 1) × 100 | The minimum volume growth at which the discount does not reduce profit. |
| Effective multibuy discount | Discount, % = (1 − Bundle price ÷ (Units × Unit price)) × 100 | The “Multibuy” calculation. |
| Points liability | Liability = Unredeemed points × Point value × (1 − Breakage share) | The breakage share is an estimate from history, not a guaranteed value. |
Formulas in notation
- Promotion price
P_1 = P_0 · (1 − d)
P_0 is the regular price excluding VAT, d is the discount depth.- Profit per unit
u = (1 − r) · (P − C) − r · c_r
r is the return share, C is cost, c_r is the loss on a returned unit (logistics, handling); promotion and baseline have their own r.- Required uplift
L = u_0 / u_1 − 1
u_0 is profit per unit outside the promotion, u_1 is within it; if L exceeds attainable uplift the promotion is loss-making.
Worked example
Promotion result: 20% off for 2 weeks (matches the platform calculation) (an independent example: prices excluding VAT RUB 5,000, on the shelf RUB 6,100 including 22% VAT)
| Item | Value |
|---|---|
| Promo price, RUB | 4,000 |
| Unit profit in normal time u0, RUB | 2,736 |
| Unit profit in promotion u1, RUB | 1,724 |
| Baseline for the promotion, units | 80 |
| Sold in promotion, units | 190 |
| Profit in promotion, RUB | 327,560 |
| Profit without promotion, RUB | 218,880 |
| Cannibalization, RUB | 27,360 |
| Pull-forward, RUB | 32,832 |
| Promotion costs, RUB | 20,000 |
| Incremental profit, RUB | 28,488 |
Sales grew by 138% over baseline, and the promotion pays back at an uplift of at least 59% even without costs. After cannibalization, pull-forward and costs RUB 28,488 remains: the discount paid back, but with a thin margin. Ignoring cannibalization and the dip would have overstated the result by RUB 60192.
Figures
- positive
- negative
- total
Data as a table
| Step | Change | Level |
|---|---|---|
| Profit during the promotion | +328k | 328k |
| Baseline profit | −219k | 109k |
| Cannibalization | −27.4k | 81.3k |
| Demand pull-forward | −32.8k | 48.5k |
| Promotion costs | −20k | 28.5k |
| Incremental profit | — | 28.5k |
Promotion profit: what is left after baseline sales and costs. Profit during the promotion minus the profit that would have come anyway, minus cannibalization, demand pull-forward and promotion costs; the remainder is the true increment.
Data as a table
| Measure | Required sales uplift over baseline |
|---|---|
| 5% | 15 |
| 10% | 26 |
| 15% | 41 |
| 20% | 59 |
| 25% | 82 |
| 30% | 113 |
What sales uplift a discount depth needs to pay off. The deeper the discount, the faster the required uplift grows: at 30% the required uplift is several times that at 10%; the curve uses the promotion return rate as in the example.
Specifics by category
- Apparel
- Elasticity is higher for fashion and seasonal items: a discount moves volume more but trains shoppers to wait. For basics the discount gives less uplift and cannibalizes more; promo is placed on end-of-season stock.
- Footwear
- A discount moves size remaining stock poorly: extreme sizes do not sell even at a discount, and size returns in promotions are higher. Prefer mechanics for specific sizes and transfers between stores; use a return share above normal in the calculation.
- Accessories
- Gift demand: “set”, “gift with purchase” and multibuy mechanics work better than deep discounts; margin is higher so deeper cuts are affordable, but image items are kept out of discounts.
- Kids
- Promo is tied to the school calendar and holidays; parents buy ahead for growth, so pull-forward is visible. Discounts on uniform in August-September and New Year sets are planned early; requirements for kids goods listings and ads are stricter.
- Men
- The audience responds to price and a clear mechanic rather than frequent promotions: two or three big events a year work better than constant discounts; a lot of gift demand (February, December).
- Women
- The most promotion-heavy category: the risk of a “wait for discounts” habit is highest, so a reference price and a frequency rule are needed. Multibuy and capsule sets give a better basket than deep discounts.
Common mistakes
- A promotion is judged by revenue growth → it looks successful while profit falls → judge incremental profit.
- Cannibalization and the post-promo dip are not counted → the result is overstated → include both corrections.
- A promo code goes to everyone → most buyers would have bought anyway → compute the self-selection share and the payback threshold.
- Stacking mechanics without rules → the final discount exceeds the plan and goes below margin → set the order of application and a total discount ceiling.
- Loyalty points are not treated as a liability → costs are hidden → provision for unredeemed points.
- The promotion result is computed against last year without a control group → the effect of weather and season is credited to the promotion → compare with a control group and baseline.
Russian specifics
- Online promotion ads require marking and an identifier (erid); the advertising law sets liability for violations. Rule: Law 38-FZ of 13.03.2006: Federal Law of 13.03.2006 No. 38-FZ on advertising (rule check not completed).
- Reducing a partner item price on a platform is possible by contract and with at least five working days notice to the partner (platform economy law). Rule: Law 289-FZ of 31.07.2025: Federal Law of 31.07.2025 No. 289-FZ on the platform economy.
- The price on the tag must match the price at the till; the retail sale rules require price tags with the unit price. Rule: Decree 657 of 30.05.2026: Government Decree of 30.05.2026 No. 657 (rules of retail sale of goods).