A post-purchase upsell discount strategy is not “always 20% off.” You are trading take rate for margin. Ten to twenty percent, a save-$ amount, or no discount can all win — depending on the pair, the cost, and who just paid. Use the table, then check profit, not vanity AOV.
Quick answer: percentage (often 10–20%) when the add-on is a cheap complement and you need a reason to tap. Fixed save-$ when you want a concrete number on a low price, or a cap on a high price. No discount when the pair is obvious, stock is tight, or a slash would cheapen a premium SKU. Never pick a discount before the product pair. Relevance first — then price.
You have three honest levers after checkout. A percent off the add-on. A fixed dollar off. Or the shelf price, with one-tap convenience as the only perk.
They are not “good, better, best.” A 20% cut on a $8 sock is $1.60. The same 20% on an $80 bottle is $16. The buyer feels those differently. Your COGS does too.
The native post-purchase page adds a line to the order they already placed. Apps attach an explicit discount to that line. You are not handing out a sitewide code. If you need the flow itself, start with one-click after checkout.
Take rate is accepts ÷ eligible views. Margin is what you keep after COGS, the discount, card fees, pick-and-pack, extra shipping, returns, and a slice of app cost.
A higher take rate with a fatter discount can make less money. A lower take rate at full price can make more. You will not know until you plug real numbers into true upsell profit.
Worked sketch (illustrative, not a promise): add-on list $20, COGS $6. At 0% off, buyer pays $20. At 15% off, buyer pays $17. If 15% only lifts accepts a little, you gave away $3 for nothing. If it lifts a lot and the extra yeses still cover $3 plus fees, keep it. That is a test, not a slogan. How to run the test: post-purchase A/B testing.
The free upsell revenue calculator multiplies orders × take rate × offer price. That is extra sales, not extra profit. The AOV calculator is average order value only. Discount strategy lives in the profit worksheet.
| Use this | When | Watch out |
|---|---|---|
| 10% off | Obvious pair, you only need a small nudge. First-time complement with healthy margin. | May be invisible on a $6 item ($0.60). |
| 15–20% off | The pair is good but not obvious. You can still keep profit after COGS and fees. | 20% on thin margin is a hobby, not a strategy. |
| Save $X | Low-priced add-on (“Save $4”) or you must cap the giveaway on a high SKU (“Save $10, not 25%”). | $X can look stingy on a $90 item or huge on a $9 item. Recheck the share of list price. |
| No discount | Must-have accessory, limited stock, premium brand, VIP already discounted, returning refill they expect to pay for. | If take rate is near zero, the pair may be wrong — not the missing coupon. Fix product rules first. |
First-time vs returning often needs two prices: a small percent for new buyers, shelf price for refills. Split: first-time vs returning.
Use a percentage when the add-on price is mid-range and you want the discount to scale with the SKU. Ten percent is a nod. Twenty percent is a sale. Stay in that band unless you have run the profit sheet and the extra yeses still pay.
Percentages feel familiar. Shoppers compare them to the PDP. If the PDP already shows 20% off, another 20% on the thank-you flow looks like the real price was fake. Either show the sale price as the list, or do not double-cut.
High-ticket add-ons: 20% can be a large dollar hole. Those carts often need a smaller percent or a capped save-$. See high-ticket post-purchase upsells.
Use a fixed amount when you want the savings to read like a price tag: “Add for $12 (save $4).” That is clearer than “20% off” on a $16 item.
Use it also as a ceiling. “Save $8” on a $40 add-on is 20%. The same “save $8” on an $80 add-on is 10%. You chose the dollar you can afford, not a percent that grows with the catalog.
A common rule of thumb: under about $100, a percentage sounds bigger (“20% off” beats “$4 off” on a $20 item); over $100, dollars sound bigger (“save $30” beats “12% off”). Treat it as a starting guess to test, not a law.
Write the final price in the button. People tap a number, not a story. Copy help: how to write post-purchase offers.
Skip the cut when the add-on is the thing they would have needed anyway: a lid, a filter, a spare, a size they already wear. One tap into the same shipment is the deal.
Skip it when stock is scarce. Discounting a rare SKU trains people to wait. Skip it for tagged wholesale or VIP if their price list is already lower. Skip it on premium lines if a red badge would cheapen the brand.
If full price gets almost no taps, do not jump to 40% off. Change the product. The usual miss is an irrelevant SKU — see post-purchase upsell mistakes.
They offered paper filters after every bag. First version: 20% off. Takes were fine. Profit was thin because filters were already a traffic-builder at almost cost.
They switched first-time buyers to 10% and returning buyers to full price with the line “We’ll drop them in the same box.” Filter profit rose. Bag sales did not fall. The pair was always the product. The discount was the leftover knob.
This trips people up every season, so here is the plumbing.
Source for the admin-discount limit: Shopify discounts FAQ.
Offering another unit of what they just bought is its own animal. Some vendors tell you to cut at least 20% here, and the logic is fair: the shopper just saw the full price ten seconds ago, so a token 5% reads as noise.
Our version: a second unit deserves a deeper cut than a complement, because your marginal cost is low (same box, same pick) and the buyer's reference price is fresh. But run the sheet first. Twenty percent off a thin-margin hero is still a hole, weekend or weekday.
If the second unit does not cover a 20% cut, offer a cheaper consumable instead — the discount question disappears when the pair changes.
Discount bugs are quiet. They show up on the order as a smaller number, not as an error toast.
Our app. One-click post-purchase, thank-you blocks, cart drawer, gifts, BOGO, and volume discounts in one bill from $9.99/mo. 14-day trial. 5.0 from 36 reviews (small sample). Built for Shopify. Watch that drawer discounts and post-purchase cuts do not stack by accident.
Compared in best post-purchase upsell apps and Oxify vs other apps.
Only if the extra yeses still leave more profit after COGS, payment fees, pick-and-pack, and returns. A 15% off offer that doubles takes can still lose if the add-on was thin margin. Run the profit formula, then pick percentage, a save-$ amount, or full price.
It is a common starting band for a small complement, not a law. Ten percent is a nudge. Twenty percent is a sale. Neither replaces a relevant product. If 20% wipes the margin, use a smaller percent, a fixed dollar cap, or no discount.
When you want the savings to look concrete on a cheap add-on (Save $4) or when you must cap the giveaway on a pricey SKU. Percentage grows with price. A fixed amount does not. Pick the one that matches how you talk about the product.
When the add-on is an obvious fit, inventory is tight, the customer is a tagged VIP who already gets a better price, or the item is premium and a slash would cheapen it. Convenience and one tap are the offer. The price can stay the shelf price.
Not always, and even when it does, profit can fall. Test discount after you have a relevant pair. Measure profit per view, not only accept rate. Stacked cart discounts plus a post-purchase cut can also double-dip — check the final line on a test order.
The native one-click flow adds a line to the existing order. Apps apply an explicit discount on that line. Do not send people to a second checkout with a code. Codes belong in email or the cart, not as a workaround for the post-purchase slot.
Vendor guides cluster around 10–20% for complements and 20% or more when re-offering the same product the customer just bought. Those are take-rate tips, not profit tips. Start at 10–15% on a complement, deeper only on a second unit with healthy margin, and confirm with profit per view.
Admin discounts do not apply on the post-purchase page, but an app-level percent can cut a product that is already marked down on the storefront. Place a test order during a sale and read the added line. If it double-cuts, exclude sale collections from the offer or set the discount to zero.
Lock the pair. Choose 10–20%, save-$, or none. Place a test order. If the extra yeses do not cover the cut, the discount is the problem — or the product is.