Start With the Business Goal
The first question before building a promotion is not "what discount should I offer?" It's "what am I trying to achieve?"
Different goals require different promotion types:
- Increase average order value — bulk pricing tiers, cart discount threshold, BXGY cross-sell
- Clear slow-moving inventory — simple discount on specific products, BXGX to move quantity
- Acquire new customers — first-order cart discount, introductory promotion
- Increase repeat purchases — post-purchase order reward coupon
- Move related products — Buy X Get Y cross-sell with a slow mover as the reward
Running a promotion without a clear goal produces campaigns that are hard to evaluate and harder to improve.
The Margin Math You Need Before Every Campaign
Three numbers matter before setting any discount:
- Product price — what the customer pays
- Product cost — what it costs you (COGS)
- Gross margin % — (price − cost) ÷ price × 100
A product priced at $50 with a cost of $30 has a 40% gross margin. If you offer 20% off, the effective price is $40. Gross margin on the promoted sale: ($40 − $30) ÷ $40 = 25%. You're still profitable, but margin has dropped from 40% to 25%.
At 40% off ($30 price, $30 cost), gross margin is 0%. At any deeper discount, you're selling below cost.
Store gross margin on each product and check it before setting any discount percentage. This is basic math — but it's skipped often enough that many promotions run below cost unintentionally.
Targeting: Why "Broad" Promotions Are Risky
Broad promotions — sitewide sales, category-wide discounts — are easy to set up and create the feeling of a big event. But they apply the same discount to every product, regardless of margin. A category might contain products with 50% margin and products with 10% margin. A 20% off category promotion is fine for the first group and potentially loss-making for the second.
Precise targeting costs more setup time but produces better outcomes:
- Target specific products you know have healthy margins
- Use margin range targeting (Pro) to automatically exclude below-threshold products from a broad campaign
- Use sale item exclusion (Pro) so already-discounted items don't receive additional promotional discounts
Thresholds: Making Customers Do More, Not Less
An unconditional discount — 20% off, no strings — changes nothing about customer behaviour. The customer who was going to buy $60 of product still buys $60 of product, just at 20% off. You've given away revenue without increasing order value.
A threshold discount — 20% off when you spend $80 — creates a specific target. The customer who has $60 in cart sees a reason to add $20 more. A significant proportion will do so, especially with a cart progress bar making the threshold visible.
For bulk pricing, set the first tier above the typical order quantity for that product. For cart discounts, set the threshold 20–30% above your current average order value. For BOGO, set the qualifying quantity above the typical single-session purchase quantity.
Stacking: The Compounding Risk Most Merchants Don't Notice
If two promotions can apply to the same order simultaneously, the effective discount is not the sum of the two. But the margin impact often is. A 15% bulk discount that stacks with a 20% storewide sale produces 35% off the affected products. If the product margin is 35%, you've broken even. If the margin is less, you're below cost.
Configure stacking controls on your promotion engine:
- First-matched — only the first matching promotion applies
- Highest-discount — only the promotion with the largest discount applies
- All-stack — all matching promotions compound
Most stores should use first-matched or highest-discount for general campaigns, and only allow stacking for intentionally combined offers.
The Offer Safety Score Framework
Ezhance's Offer Safety Score evaluates six dimensions before you publish any promotion:
- Margin Safety (35 pts) — price-weighted gross margin of targeted products
- Stacking Protection (20 pts) — whether multiple promotions can compound
- Purchase Threshold (15 pts) — whether there's a minimum quantity or value signal
- Abuse Control (10 pts) — whether customer targeting limits who qualifies
- Offer Clarity (10 pts) — whether display settings include a title, description, and urgency signal
- Data Confidence (10 pts) — how specific the targeting is
A score above 70 indicates a well-structured promotion. A score below 50 indicates a promotion with multiple unaddressed risks. Reviewing the score before publishing catches problems that are expensive to discover post-launch.
Measuring Whether It Worked
After a promotion runs, two numbers tell you whether it was profitable. Check them in the Ezhance analytics dashboard:
- Revenue attributed to the promotion — total order value from orders where the promotion applied
- Total discount given — the aggregate discount amount across all orders
If revenue attributed to the promotion increased compared to the same period without a promotion, and the total discount given is less than the incremental revenue, the promotion generated net-positive return.
If revenue didn't increase, the promotion didn't change customer behaviour — it just reduced revenue on purchases that would have happened anyway.