Cross-Sell Email Examples That Convert

A cross-sell email recommends a complementary product to someone who has already bought. The examples that still convert share one property: they reference the specific item the customer owns and solve a problem that item created. Generic recommendation grids no longer clear their own unsubscribe cost, and the benchmark data shows why.

Start with the number the other examples leave out

Cross-sell is usually presented as free money. The current data says otherwise, and it is worth seeing before you look at any example.

Omnisend's benchmark set, built on more than 20 billion campaign emails and 470 million automated sends across 27,000 brands, puts the cross-sell automation in 2025 at a 42.09 percent open rate, a 3.02 percent click-through rate, a 0.87 percent conversion rate and 0.95 dollars per email. The unsubscribe rate is 0.89 percent, which is the highest of any automation type in that table.

Now the trend. Omnisend's 2023 report, built on 23 billion emails, put the same automation at a 46.6 percent open rate and a 2.56 percent conversion rate. Conversion has fallen by roughly two thirds in two years. The open rate barely moved, which tells you the problem is not the subject line.

For contrast, back-in-stock in the 2025 dataset runs a 58.80 percent open rate, a 21.31 percent click-through rate, a 6.72 percent conversion rate and 9.14 dollars per email. That is the same channel, the same audience, roughly eight times the conversion. The difference is that back-in-stock answers a request the customer actually made.

That is the frame for everything below. A cross-sell works when it behaves like an intent-matched email. It fails when it behaves like a catalogue.

Five examples, and why each one works

The consumable companion. A customer buys a coffee grinder; the email offers the beans, filters and a descaling sachet as one bundle, priced as a set. It works because the purchase created a need that did not exist before, and because the bundle removes a decision rather than adding one. Send it after the product has been used, not after it shipped.

The size or refill step-up. A customer buys the 30-day supplement; the email offers the 90-day pack with the per-unit saving shown. This is the most reliable cross-sell in consumables because it is not really a cross-sell: the customer has already chosen the product, and the only question is quantity. Show the per-unit price difference explicitly or the offer reads as an upsell for its own sake.

The category adjacency, tied to a stated reason. A customer buys running shorts; the email offers anti-chafe balm with one line explaining that most people who buy these shorts for long runs buy it too. The reason is what converts. The same email without the reason is a product grid.

The review-triggered recommendation. A customer leaves a four or five star review; the email thanks them and recommends the one product most often bought next by customers who rated that item highly. This performs because the customer has just told you they are satisfied, which is the only moment at which a recommendation is unambiguously welcome. It also links your reviews programme directly to revenue, which most brands never do.

The gift-lane cross-sell. A customer buys something that is plausibly a gift; the email offers a card, wrap or a small add-on. Narrow, seasonal, and the highest-margin of the five when the timing is right.

Notice what is absent: the "you might also like" grid with four unrelated products. That is the format whose conversion halved, and it is the format most cross-sell examples still show.

The three placements that survive the arithmetic

Given a 0.89 percent unsubscribe rate, a cross-sell has to earn its place. Three placements do.

After the review request, inside the post-purchase flow. The customer has judged the product, so a recommendation is credible. This is the default position and the one most brands get wrong by placing cross-sell before the review.

Inside the order confirmation, as a single secondary block. Order confirmations averaged a 53.99 percent open rate in 2025 and generated 1.60 dollars per send in Omnisend's 2026 report, so the attention is already there and the marginal cost of one block is near zero. One block, not a carousel.

Attached to a replenishment reminder. The customer is reordering anyway, so the cross-sell rides on an email they wanted. This is the only placement where a cross-sell costs you nothing in unsubscribes, because the email was going to send regardless.

Everywhere else, be sceptical. A standalone cross-sell campaign to the whole list is the version of this tactic whose numbers collapsed, and it is still the most commonly sent.

The reason those three work and the standalone does not comes down to where the attention was already going. Omnisend's 2026 report, covering 27 billion emails from 150,000 brands, found automated emails made up just 2 percent of sends while driving 30 percent of revenue, earning 2.87 dollars per send against 0.18 dollars for scheduled campaigns. A cross-sell placed inside an automation inherits that attention. A cross-sell sent as its own campaign has to earn it from nothing, against a customer who did not ask.

Subject lines for each of the five

Subject lines for cross-sell fail in a predictable way: they lead with the product being sold rather than the product already owned. Reverse that and the open rate follows.

For the consumable companion, name the purchased item and the gap: "Your grinder needs beans". For the size step-up, put the arithmetic in the subject: "The 90-day pack works out cheaper". For the category adjacency, borrow the reason: "Most people who bought those shorts also bought this". For the review-triggered recommendation, acknowledge the review first: "Thanks for the review. One more thing you might like". For the gift lane, be literal: "Need a card with that?"

Two rules across all of them. Keep the owned product inside the first thirty characters so it survives mobile truncation. And do not use urgency, because there is no deadline and the customer knows it, which makes a false one expensive on a list you are already asking a lot of.

Preview text should do separate work rather than repeating the subject. Use it for the reason: if the subject names the owned product, the preview names the problem the recommendation solves. That pairing is what lifts click-through on this flow, and it is the change that takes ten minutes.

What to measure, and the metric to lead with

Lead with unsubscribe rate, not conversion. That inverts how most teams read this flow and it is the correct order, because a cross-sell that converts at 1 percent while shedding subscribers at 0.9 percent is trading a durable asset for a one-off order.

Set a ceiling before you launch. If the cross-sell's unsubscribe rate runs above your account's flow average, the targeting is too broad regardless of what revenue it books. Narrow the audience to a category the customer has already engaged with, or cut the email.

Then measure incrementally. Hold the cross-sell back from 10 percent of eligible customers for a full cycle and compare total revenue per customer, not cross-sell revenue. Cross-sell revenue always looks positive in isolation, because it counts orders that would have happened anyway through a different route.

Our retention email programmes team sets the audience rule and the unsubscribe ceiling before anything is written.

Copy rules that hold across all five examples

Name the product they bought, in the first line. "Because you picked up the Aeropress" outperforms any variation that opens with the recommended item, because it establishes that this email is about them rather than about your catalogue.

Offer one product, or one bundle. Two is a comparison, four is a grid, and a grid is a browsing experience delivered to someone who is not browsing.

Give the reason in one sentence, before the price. The reason is doing the conversion work.

Keep the discount out of it unless the bundle needs one to make sense. A cross-sell with a discount attached trains customers to expect a discount on every recommendation, and the margin loss compounds across the whole flow.

Reference the reason in the body, not only in the subject. A subject line that promises a reason and a body that shows a product grid is the most common mismatch in this flow, and it is why click-to-conversion on cross-sell runs so far below back-in-stock.

Do not send the same cross-sell twice. If it did not work, the recommendation was wrong, and repeating it makes the unsubscribe rate worse without improving the conversion rate.

Where the recommendation depends on review content to be credible, the reviews programme has to be producing that content first. That is the dependency most cross-sell builds ignore, and it sits with the reviews and UGC programme. Where the cross-sell rides on a reorder, it belongs to the subscription and reorder path rather than to the campaign calendar.

In-body table: Cross-sell against the alternatives, 2025 Omnisend automation data