How to apply upsells and cross-sells

Three placements with three rules: PDP cross-sells, cart upsells, post-purchase email — including plant-specific bundle ideas.

Written By Bas den Hoed

Upsells and cross-sells are the most over-used and worst-tuned feature in plant e-commerce. Most reseller shops drop the same "you may also like" widget on every PDP, in the cart, and in the post-purchase email — same logic, same products, three placements. Customers ignore all three. This article is the placement-specific playbook: three touchpoints, three different jobs.

💡 The most common reseller mistake: stuffing every upsell on the product detail page. The PDP, the cart, and the post-purchase email each ask the customer a different question. Showing the same six "bestsellers" everywhere answers none of them.

Three placements, three different jobs

Each touchpoint has a different customer mindset. The upsell strategy follows the mindset, not the inventory.

PDP cross-sell — "Complete the look"

The customer hasn't decided yet. The cross-sell job is to show what completes the plant — a matching pot, a complementary plant, soil and a watering can if it's a beginner species. The mood is "imagine this fully styled".

  • Maximum 4 items. More than four turns the cross-sell into a second product list and dilutes the page's primary CTA.

  • Hand-curated for top sellers, algorithmic for the long tail. Your top 20 deserve a human-picked combination. Below that, an algorithm based on category and price tier is fine.

Cart upsell — "Customers also added"

The customer is committed to the main item. The cart-upsell job is to add small, high-margin extras that fit the cart — not to introduce a second decision. The mood is "one more click, not one more thought".

  • Maximum 2–3 items. A long list here is friction, not revenue.

  • Price each upsell around 15–30% of the cart total. A €30 plant + a €6 saucer feels natural; a €40 designer pot reads as a second purchase decision.

  • "Add with one click", not "View product". Sending the customer back to a product page resets the journey.

Post-purchase email — "Care kit for your new plant"

The customer has bought; the plant is on its way or has arrived. The post-purchase job is timing-and-relevance: a watering can two days after delivery, fertiliser two weeks in. The mood is "we're still here after the sale".

  • Maximum 1–2 items per email. One specific recommendation outperforms a list.

  • Triggered by event. 5 days post-delivery for accessories, 14 days for follow-on plants in the same care category, 60 days for a "how's your Calathea?" check-in.

  • This is the highest-converting placement when done well — the customer already trusts you, they have the plant in front of them. Don't waste it on a generic newsletter.

Plant-specific bundle ideas

Bundles work when the items belong together in the customer's head. Three patterns that consistently land in plant retail:

  • Plant + matching pot. The simplest bundle and the most visually obvious. Hand-pick the pot — algorithmic matches usually clash. Best displayed as a styled lifestyle shot of the combination.

  • Plant + soil + watering can. The "starter kit" for first-time buyers. Works especially well for low-maintenance plants (Sansevieria, ZZ, Pothos) where the customer is signalling they want a complete solution.

  • Two plants for the same light condition. "Both thrive in low light" or "Both love bright indirect" — the customer's space gets one decision (light condition) instead of two product decisions. Particularly effective for customers buying for a specific room.

Avoid the obvious bad pattern: pairing a plant with an unrelated bestseller because the algorithm couldn't find anything closer. A Calathea Orbifolia next to a Christmas tree in the cross-sell widget is a tell that the system is broken — better to show nothing than noise.

Hand-curated vs algorithmic

  • Hand-curate for your top 20 sellers, seasonal heroes, and plant + accessory bundles where styling matters.

  • Algorithmic for the long tail and cart upsells. Tune around category + light requirement + price tier, not just "frequently bought together".

  • Audit monthly. Open your top 20 PDPs as a customer. If the cross-sell looks random, fix it — the algorithm doesn't audit itself.

For the spec data that drives algorithmic matches — light requirement, mature size, care level — see product specifications. Better specs in, better matches out.

Bundle pricing — 5–10%, no more

A 5–10% discount versus separate purchase is enough to nudge the customer toward the bundle without eating margin. Above 10% you train customers to wait for bundles; below 5% the discount isn't visible enough to motivate. Show the saving explicitly: "Buy together: €34.95 (save €3.50)". Vague "bundle deal" framing without the number doesn't move the needle.

⚠️ If the algorithm returns nothing relevant, show nothing. A blank cross-sell area is better than four random products that signal "we're not paying attention". This is the single config most reseller shops have wrong.

Next action

Open your three best-selling PDPs as a customer. Ignore everything except the cross-sell widget. Are the four items genuinely "complete the look", or bestsellers from unrelated categories? Check the cart with one of those products in it — same audit. Then send yourself a post-purchase email by ordering a plant. Most reseller shops fail one of these three checks; many fail all three. Fix the worst-performing touchpoint first. See also pricing and stock for bundle-price display.