Upsell Pricing: How Much Discount Is Enough

An upsell discount has an odd property: it costs you nothing on the orders it persuades and something on every order that would have added the item anyway. Everything about pricing an add-on follows from that. The question is not whether a discount makes the add-on more attractive. It does. The question is how many extra attaches you need to cover the discount you also hand to the shoppers who were already buying.

This page derives that number, gives a lookup table for common margins, and ends with the cases where the right discount is none. All figures are assumptions for arithmetic, not benchmarks. No attach rate in this post comes from a study.

What a discount really costs on an add-on

Define the attach rate, b, as the share of orders containing the base product that also contain the add-on. If 6 of every 100 base-product orders include the add-on today, b is 6 percent. Those 6 shoppers will pay full price whether or not you discount. When you add a discount, they are the ones who pay the cost, because you now earn less on a sale you already had.

The shoppers who attach because of the discount are the gain. Each of them earns you the discounted margin on a sale you would not have had. A discount works when the gain is at least the cost. That is the whole model, and it has only three inputs: the add-on's margin, the discount and today's attach rate.

The formula, derived

Let p be the add-on's price and m its contribution margin as a decimal, after product cost and payment fees. Let D be the discount as a decimal of price, with D smaller than m. Take 100 base orders.

• Before the discount, add-on profit is 100 × b × p × m.
• After the discount, each attach earns p × (m − D), because the price falls by p × D and the cost does not. If the attach rate rises to b′, profit is 100 × b′ × p × (m − D).
• Break even means the two are equal: b′ × (m − D) = b × m.

So the attach rate you need is b′ = b × m ÷ (m − D). The multiple m ÷ (m − D) is the lift required, and it does not depend on the price p, which cancels. The extra attaches needed are b × D ÷ (m − D) per base order.

This is not the whole-order result. When a discount applies to every order, the cannibalized base is all orders, and the table in Break-Even Discount: Extra Volume You Need covers that case. Here the cannibalized base is only the share that attaches today, which is why the same percentage can be affordable on an add-on and costly store-wide.

Lookup table: the attach rate multiple you need

Each cell is m ÷ (m − D). Read it as the factor by which today's attach rate must rise: 1.50 means it must go up by half.

Two patterns matter. A thin margin turns a modest discount into a steep climb: 20 percent off a 30 percent margin add-on needs the attach rate to triple. And a fat margin forgives: at 70 percent, even 25 percent off needs only a lift of about 56 percent.

A worked example

Assume an add-on that sells for 24 dollars and costs 9 dollars to buy, ship in its slot and process, so the margin is 15 ÷ 24 = 62.5 percent. Assume 6 percent of base orders attach it today. Before any discount, 100 base orders earn 6 × 15.00 = 90 dollars from the add-on.

Each rate in the fourth column is 90 ÷ profit per attach. A 15 percent discount has to lift the attach rate from 6.00 to 7.89 percent, about 32 percent more attaches, just to leave you where you were. Anything above that is profit. Anything below is a gift to the shoppers who were already buying. If you cannot see a plausible path to that lift on your own product page, take the smaller discount.

How to find b: count orders last month that contained the base product, count how many of those also contained the add-on, and divide. Do it in a normal month with no offer running, so the number is a baseline and not a result.

When the attach rate is already high

The same formula warns you off discounting the add-ons that already sell well. Take the same 62.5 percent margin and a 10 percent discount, but start from an attach rate of 40 percent. The extra attaches needed are b × D ÷ (m − D) = 0.40 × 0.10 ÷ 0.525, which is 7.6 percentage points, compared with 1.14 points when the base rate was 6 percent. Every point of baseline attach rate you already have is a point of full-price sales you are about to mark down.

No discount at all

Some add-ons sell on convenience: the matching case, the refill, the extra charger. If shoppers want the item because it fits what they are buying, a price cut transfers money without changing behavior. Test the add-on with no discount first, show it at the moment of need, and compare the attach rate with your baseline. Add a discount only if the result is below what the margin can afford, and use the table to see how far the rate has to move.

Discount the cheaper item of the pair

A percentage costs less in dollars on a smaller price. Suppose a 60 dollar anchor product and a 20 dollar add-on. Fifteen percent off the add-on costs 3.00 dollars when it is taken. Fifteen percent off the whole 80 dollar pair would cost 12.00 dollars. If the offer is aimed at the add-on, discount the add-on and leave the anchor at full price.

Shopify's own mechanics point the same way. Its Buy X get Y help page says that when the product or collection chosen as the discounted item is the same as the one in the Customer buys section, the lower-priced item the customer chooses is the item discounted. Its amount-off discount page says a discount can apply to a specific product, a collection or an entire order. Choose the narrow target, the add-on, so the discount cannot leak onto the anchor.

Margins you can trust

The table is only as good as m. Use contribution margin: price minus product cost, minus the payment fee on that sale, minus anything you pay to pack or ship the add-on. Your supplier invoices and your processor statements are the source, and the method for entering costs so reports can use them is in Discount Floors So a Promo Never Loses Money. If margin is unknown, the table cannot help you. For a deeper look at why percentage discounts cost more than they appear to, see Why a 15% Discount Costs More Than You Think.

Build the offer

Pick one add-on, choose the smallest discount from the table whose required lift you can believe, and show it on the product page of the anchor. The Frequently bought together template in Freebies Engine by Log Studio Lab suggests matching products on the product page, and the setup choices for pairs and placement are in Frequently Bought Together on Shopify. A checkout version of an upsell exists, but it requires Shopify Plus. Measure the attach rate for a full normal month, then compare it with the number in the table. Average order value alone will not tell you whether the discount paid for itself, as Average Order Value: Formula and 12 Levers explains.