TL;DR
- Cross selling adds a complementary item to the cart; up selling swaps the item for a higher model; bundling sells a priced set.
- Market basket analysis picks pairs: lift above 1 means buyers of A choose B more than average; lift near 1 means the pair is bought together no more often than chance.
- Measure attach rate and revenue per visitor in an A/B test, not AOV alone, and watch checkout completion for lost orders.
- Using purchase history for marketing messages in Thailand needs a lawful basis under Thailand's 2019 Personal Data Protection Act (PDPA), such as recorded consent.
Cross selling is offering another product or service that goes with what the customer is already buying, such as a case and screen protector for someone buying a phone. The goal is for the customer to get everything they will need in one purchase and for the shop to earn more per order. It differs from up selling, which invites the customer to switch to a higher-priced version instead of the original item.
This guide covers how cross selling works compared with up selling, illustrative examples for online shops, how to run it through a website, email and LINE OA, how to use purchase data to choose what to offer, the metrics that matter such as AOV and attach rate, and how to avoid annoying customers.
What cross selling is and how it differs from up selling
Both aim to raise the value of a purchase, but they point in different directions. Cross selling adds a new related item to the cart. Up selling invites the customer to choose a more expensive model or bigger package in place of the one they meant to buy. A third method is often grouped with them: bundling, which sells several items as a set at one price.
| Method | What is offered | Illustrative example |
|---|---|---|
| Cross selling | A complementary item added to the cart | Buy a camera, get offered a memory card and bag |
| Up selling | A higher model or package replacing the original | View the standard camera, get offered the model with sharper video |
| Bundling | Several items as one priced set | A camera kit with lens, memory card and bag |
In practice most shops use all three. The difference to remember is that a good cross sell solves a problem the customer will hit after buying, such as a printer with no spare ink, while a good up sell has to explain why the pricier model is worth it for this particular customer's use.
Why cross selling works
The main mechanism is timing. Once a customer has decided to buy, they are thinking about using the product. Offering what goes with it at that moment saves them coming back to find it later, and saves the shop paying for ads to bring the same customer back for a small item.
The other reason is shipping. In online retail, shipping cost per order usually does not rise in direct proportion to item count, so a small item added to an existing order often costs less to deliver per item than shipping it separately. This is why many shops set a free-shipping threshold and recommend products that help customers reach it.
Illustrative cross sell and upsell examples for online shops
Every example in this section is invented to explain the method. The figures are not data from a real shop.
- Skincare shop. A customer adds a serum to the cart. The cart page shows a morning sunscreen that pairs with it, with a short note on the order of use. That is a cross sell. Offering the larger serum size with a lower price per millilitre is an up sell.
- Pet supplies shop. A customer buys cat litter. The system offers a litter scoop and odour-control bags, which are used with it every day.
- Online coffee shop. A customer buys roasted beans. The order confirmation page offers filter papers that fit the dripper the customer bought before, because the shop has their purchase history.
- Subscription software. A customer on the basic plan is close to their storage limit. The system offers an add-on for extra storage, which is a cross sell of an add-on. Offering a move to the whole higher plan is an up sell.
Now an illustrative calculation. A shop has 1,000 orders a month with an average of 600 baht. If 15% of orders add a 150-baht accessory, revenue rises by 150 orders x 150 baht = 22,500 baht, and average order value moves from 600 to 622.50 baht. The numbers only show how to calculate it; real results depend on each shop's products and customers.
Where cross selling fits in the purchase journey
Where the offer appears affects both sales and annoyance, and each point suits a different kind of offer.
- Product page. Suits a "frequently bought together" or "goes with" block the customer can add in one click. Place it below the buy button, not ahead of the main product information.
- Cart. Suits small, low-priced items that help reach the free-shipping threshold. Customers decide quickly because the price is low.
- Checkout. Use with care. Every step added to checkout is a point where the customer may leave before paying. If you offer anything, make it a single optional choice that does not change the page.
- After purchase. The order confirmation page, email and LINE OA messages suit offers that need time to consider, or items the customer will need when the main product runs out.
Running cross selling through website, email and LINE OA
On the website
Most e-commerce platforms, such as Shopify and WooCommerce, have built-in features or plugins for related products. They usually work in two modes: manual per product, which gives good control but takes time with a large catalogue, and automatic based on purchase history, which needs enough order data to be accurate. A new shop can start by setting pairs manually for its best sellers.
By email
Post-purchase emails carry order and shipping details customers want to check, which makes them a natural place for a small recommendation. A common sequence is an order confirmation with a small recommendations section at the bottom, a how-to-use email after delivery that suggests items that improve use, and a replenishment reminder when a consumable is about to run out, timed from pack size and approximate usage rate.
Through LINE OA
Many Thai customers talk to shops on LINE, so it is an important channel. But LINE OA messages carry quotas and costs under each plan, and customers can block an account easily, so sending the same message to every follower does not suit cross selling. A better approach is to connect purchase data to a LINE CRM, segment by what customers have bought, send only offers related to what they own, and use a Rich Menu so customers can browse add-ons themselves.
In marketplace stores
Shopee and Lazada offer promotion tools in their seller centres, such as product bundles and discounts on add-on purchases. Tool names and conditions can change with platform policy, so check the seller centre before planning a campaign.
Using purchase data to choose what to offer
The widely used method is market basket analysis: finding which product pairs are bought in the same order often. It uses three numbers.
- Support. The share of all orders that contain both A and B.
- Confidence. Of orders containing A, the percentage that also contain B.
- Lift. Confidence divided by the share of all orders containing B. Lift above 1 means people who buy A are more likely to buy B than customers in general. Lift near 1 means A and B appear together about as often as chance would predict, so buying A says little about buying B; this is typical of a best seller that turns up in many carts anyway.
Illustrative example: out of 2,000 orders, 400 contain the serum, and 120 of those also contain sunscreen. Confidence is 120/400 = 30%. If sunscreen appears in 300 of all orders, or 15%, lift is 30%/15% = 2. People buying the serum are twice as likely as the average customer to buy sunscreen, so the pair is worth testing as a recommendation.
For small shops with few orders these numbers swing around, so lean on knowledge of how products are used and confirm with data as orders build up. Watch personal data rules too: using purchase history to send marketing messages needs a lawful basis under Thailand's Personal Data Protection Act B.E. 2562 (2019), such as properly recorded consent.
Measuring cross selling with AOV and attach rate
| Metric | How to calculate |
|---|---|
| AOV (average order value) | Total revenue divided by number of orders |
| Attach rate | Orders with the add-on divided by orders with that main product |
| Recommendation click rate | Clicks divided by times the recommendation block was shown |
| Share of revenue from recommendations | Revenue from items added via the block divided by total revenue |
AOV gives the overall picture but moves for many reasons, such as promotions or price changes, so track the attach rate of the pairs you are recommending as well. Also watch add-to-cart rate and checkout completion alongside it. If AOV rises while checkout completion falls, the offer may be getting in the way and costing orders.
The clearest answer comes from an A/B test. Split visitors into two groups, one that sees the recommendation block and one that does not, and compare revenue per visitor rather than only AOV, because AOV ignores people who leave without buying. In GA4, if ecommerce events such as view_item_list, select_item, add_to_cart and purchase are set up with item data and an item_list_name for the recommendation block, you can see which products were added from recommendations.
How to avoid annoying customers
- Offer fewer, better-matched items. Two or three truly related options beat ten the customer has to scroll through.
- Keep add-on prices well below the main item. An expensive add-on forces a whole new decision and can stall the first purchase.
- Never add items to the cart automatically. Pre-ticked add-ons make customers feel tricked, and some markets ban them for extra payments outright, such as the EU under its Consumer Rights Directive.
- Do not offer what the customer already owns. Check purchase history before sending an email or message.
- Do not recommend out-of-stock items. Have the recommendation block pull only items in stock.
- Cap frequency on LINE and email. Set a weekly message ceiling per person and check block and unsubscribe rates on every send.
Common mistakes
- Recommending the shop's overall best sellers instead of items related to the cart, which usually gives lift near 1 and does not help the customer.
- Measuring only AOV without checkout completion, so you miss that the offer is losing orders.
- Setting up the recommendation block once and never revisiting it while the catalogue keeps changing.
- Discounting add-ons so heavily that the extra margin from the bigger order disappears into the discount.
Cross selling in the Thai market
Many Thai shops sell on several channels at once: their own website, Shopee, Lazada, TikTok Shop, Facebook and LINE. One customer's purchase data ends up scattered across them. Accurate cross selling starts with pulling order data from the channels the shop owns, such as the website and LINE OA, into one place, because customer data on marketplaces is usually limited by platform policy. Thai shoppers are also used to asking an admin in chat before they buy, which makes the admin a practical cross-sell point, provided admins have a ready list of recommended pairs instead of each person improvising.
If you want cross selling to work across the whole shop, see RA's e-commerce marketing, segmentation and messaging through LINE CRM, post-purchase sequences in email marketing, and recommendation blocks on product pages in e-commerce website design.
Frequently asked questions about cross selling
What is cross selling, in short?
Cross selling is offering a product that goes with what the customer is buying so they get everything in one order, such as offering a case to someone buying a phone.
What is the difference between cross selling and up selling?
Cross selling adds a new related item to the cart, while up selling invites the customer to switch to a higher model or package instead of the one they planned to buy.
What is attach rate?
Attach rate is the share of orders that include the add-on out of all orders that include the main product, and it shows how well a recommended pair performs.
Can a small shop with few orders do cross selling?
Yes, by defining product pairs manually based on how customers actually use the products, then confirming with order data once enough orders have built up.
Does using purchase history for offers breach the PDPA?
It can be lawful if there is a valid lawful basis, such as clearly recorded marketing consent, customers were told about this use in a privacy notice, and they can object to direct marketing; check with your organisation's data protection lead.
