You're running a 20% off sale next month. Should you double your inventory? Triple it? The answer depends on understanding your promotional lift—and what happens when the promotion ends.
Why Promotional Forecasting Is Hard
Promotions break all the normal demand patterns:
Demand Spikes During the Promotion
The whole point of a promotion is to increase sales. But by how much? 20%? 100%? 500%? The range of possibilities is huge.
Post-Promotion Dips
Customers who buy during a promotion aren't buying again immediately after. Post-promotion periods often see demand below baseline—customers pulled forward their purchases.
Cannibalization Effects
A promotion on Product A might reduce sales of Product B. Customers switch to the promoted item rather than adding incremental purchases.
Irregular Timing
Promotions don't follow predictable seasonal patterns. They happen when you decide they happen, based on marketing calendars, retail partner requests, or inventory situations.
Statistical forecasting methods that work well for baseline demand struggle with these dynamics.
Measuring Promotional Lift
Before you can forecast promotions, you need to understand your historical lift.
Calculate Lift Factor
Lift Factor = Promotional Period Sales / Baseline Expected Sales
Example: During your last 20% off sale, you sold 500 units. Your baseline forecast for that period would have been 200 units.
Lift Factor = 500 / 200 = 2.5x
This means your 20% discount drove 2.5 times normal demand.
Build a Lift Library
Track lift factors across promotions:
| Promotion Type | Lift Factor |
|----------------|-------------|
| 10% off | 1.3x |
| 20% off | 2.0x |
| 25% off | 2.5x |
| BOGO | 3.0x |
| Bundle deal | 1.8x |
Over time, you'll see patterns. Deeper discounts drive higher lift. Some promotion types work better than others for your products and customers.
Factors That Affect Lift
Not all promotions are equal. Lift depends on:
Discount depth: Deeper discounts typically drive higher lift—but with diminishing returns.
Promotion visibility: A promotion featured on your homepage or in email campaigns lifts more than one buried on a product page.
Product type: Price-sensitive categories see more lift than premium/luxury products.
Customer base: New customer acquisition promotions may lift differently than loyalty promotions.
Competitive context: A promotion when competitors are also promoting may lift less than during quiet periods.
Forecasting Promotional Demand
Step 1: Establish Baseline
What would demand be without the promotion? Use your standard forecasting method, adjusted for seasonality.
Step 2: Apply Lift Factor
Multiply baseline by expected lift based on similar past promotions.
Promotional Forecast = Baseline × Lift Factor
Example: Baseline forecast: 300 units × Expected lift: 2.0 = 600 units
Step 3: Adjust for Unique Factors
Is this promotion getting more or less visibility than comparable past promotions? Is the competitive environment different? Adjust the lift factor if warranted.
Step 4: Build Scenarios
Promotions are uncertain. Create low, base, and high scenarios:
- Low (75% of expected lift): 450 units
- Base (expected lift): 600 units
- High (125% of expected lift): 750 units
Plan inventory to cover at least the base case, ideally the high case for important promotions.
Accounting for Post-Promotion Dips
Promotions don't just create a spike—they often borrow from future demand.
Forward-Buying Effect
Customers who buy during a promotion buy less afterward. If you typically sell 100 units per week:
- Promotion week: 300 units (3x lift)
- Week after: 60 units (0.6x baseline)
- Two weeks after: 80 units (0.8x baseline)
- Three weeks after: 100 units (back to normal)
The total demand over a month might not change much. It just shifted into the promotional window.
Measuring the Dip
Calculate post-promotion performance the same way you calculated lift:
Post-Promotion Factor = Actual Post-Promo Sales / Baseline Expected Sales
If you consistently see a 0.7x factor for two weeks after promotions, build that into your forecast.
Planning for the Dip
Knowing a dip is coming changes your planning:
- Don't panic when post-promotion sales are slow—it's expected
- Reduce inbound orders for the post-promotion period
- Consider the full promotional window (promotion + dip) when evaluating promotion ROI
Cannibalization and Halo Effects
Promotions on one product affect others.
Cannibalization
When a promoted product steals sales from non-promoted products. Common when:
- Products are substitutes (different sizes or flavors of the same thing)
- Customers have limited budgets
- Products serve the same need
If your Vanilla flavor is on sale, Chocolate sales might drop 20%. Factor this into your Chocolate forecast.
Halo Effects
Positive spillover when a promotion drives traffic that buys other products. The 20% off sale brings customers who also buy full-price items.
Halo effects are harder to measure but worth tracking if you're evaluating promotion effectiveness.
Promotion-Specific Planning
Different promotion types require different planning approaches:
Flash Sales / Daily Deals
Short duration, high intensity.
- Stock heavily for the concentrated period
- Expect post-sale dip
- Have inventory positioned for fast fulfillment
Site-Wide Percentage Off
Longer duration, moderate lift across many products.
- Apply lift factors by product based on price sensitivity
- Watch inventory levels daily and adjust marketing if products run low
Buy One Get One (BOGO)
High lift on promoted items, significant cannibalization.
- Double your inventory for promoted items
- Reduce forecasts for substitute products
- Account for margin impact in inventory investment
Bundle Deals
Moderate lift, multiple products involved.
- Plan components together
- Consider kit production lead time if bundling at your warehouse
- Watch for imbalanced component inventory
Retail Partner Promotions
Visibility and lift depend on retailer execution.
- Get commitment details early (feature placement, circular inclusion)
- Plan conservatively—retailer promotion execution varies
- Build safety stock for retail channel specifically
Building Promotion Into Your Planning Process
Maintain a Promotional Calendar
All promotions, all channels, planned as far ahead as possible. Include:
- Promotion dates
- Expected lift by SKU
- Inventory requirements
- PO deadlines
Communicate Across Teams
Marketing knows promotion plans. Operations needs to know to plan inventory. Finance needs to know to plan cash flow. Connect these teams around the promotional calendar.
Review and Adjust Lift Factors
After each promotion, compare actual lift to forecast. Update your lift library with fresh data. Lift patterns can change as your customer base and competitive situation evolve.
Key Takeaways
- Promotional lift varies by discount depth, visibility, product type, and competitive context
- Build a library of lift factors from past promotions to improve forecast accuracy
- Account for post-promotion dips—promotions often borrow from future demand
- Consider cannibalization of non-promoted products when planning inventory
- Plan inventory using scenarios: low, base, and high lift
- Integrate promotion planning into your regular forecasting cadence
Frequently Asked Questions
Q: How much should I increase inventory for a promotion?
Apply your historical lift factor for similar promotions. If you don't have history, start with conservative estimates (1.5-2x for modest promotions, 2-3x for deep discounts) and build data.
Q: How long does the post-promotion dip last?
Typically 1-3 weeks, depending on your purchase cycle. Track your specific pattern by measuring post-promotion sales against baseline.
Q: What if the promotion performs better or worse than expected?
If better: consider cutting the promotion short before you stock out, or lean into it if you can get fast replenishment. If worse: consider extending or deepening the promotion to move inventory.
Q: How do I forecast promotions on products I've never promoted before?
Use lift data from similar products in your line. As a starting point, expect category-average lift and adjust based on product characteristics.
Q: Should I promote slow-moving inventory to clear it out?
Maybe, but be careful. Deep discounts can destroy margin. Consider whether slow movement is a forecast problem (you ordered too much) or a product problem (nobody wants it). Promotions solve the first problem, not the second.