Manual vs AI Frequently Bought Together: How to Choose

Short answer: start with hand-picked Frequently Bought Together pairs, switch the catalog to automatic (AI) pairings once your store has order history, and keep hand-picked pairs for your top sellers. We found no independent head-to-head test of manual vs AI; the only published figure is a vendor's, so measure on your own store.

Below: what Shopify's free tool already does, what that one benchmark does and does not prove, what the main options cost, and how to run a hybrid setup. Last checked: October 2026. Disclosure: we make Oxify Cart Drawer & Upsells and Oxify Quantity Breaks, both mentioned below.

The 20-second answer

• Little order history (under roughly 100 orders, a rule of thumb from Growth Suite): start manual. Automatic pairing needs purchase history to learn from.
• Past that, with a catalog too big to pair by hand: turn on automatic pairing for the catalog and check what it shows.
• Most stores: go hybrid. Let automatic pairing cover the catalog, hand-pick pairs for your top sellers, and exclude products you never want offered, such as clearance items. Shopify's own app has a setting for this.
• Everyone: give shoppers a reason to take both items (our view: a small bundle discount) and show the box on the product page, in the cart, and after checkout.

What is Frequently Bought Together?

Frequently Bought Together (FBT) is the "buy these together" box on a product page. It shows 2 or 3 items that go with the one the shopper is viewing, plus a one-click "Add all to cart" button, often at a bundle price. Amazon made the format familiar. Its job is to turn a one-item order into a two or three-item order by pre-picking the add-ons.

There are two ways to fill the box. You pick the items yourself (manual), or software picks them from your sales data (automatic, often marketed as AI). Here is how each works.

Manual FBT: you pick the pairings

With manual FBT, you hand-pick which products show together. You decide the phone case goes with the phone, and the socks go with the shoes.

The good:

• Full control. Every pairing matches your brand.
• Works on day one. No order history needed.
• Great for curated brands, kits, and obvious combos like skincare routines or full outfits.

The bad:

• It does not scale. A 200-product store means hundreds of pairings by hand.
• It goes stale. You have to update it as your catalog changes.
• It is a guess. You think two items pair well, but real buyers may disagree.

AI FBT: the data picks the pairings

With AI FBT, software reads your past orders and finds which products people actually buy together. Then it shows those pairs automatically.

Here is a simple example of how it thinks (the numbers are illustrative). Say 200 of your orders include "running shoes." Of those, 40 also include "performance socks" and 25 include a "shoe cleaning kit." The system spots both pairings and ranks socks higher, because more people bought them together. As new orders come in, it recalculates. A weak pairing at 100 orders can become a strong one at 300.

The good:

• Scales on its own. Add a product and the system folds it in.
• Finds hidden gems you would never guess by hand.
• Updates itself as buying habits change.

The bad:

• It needs data. With few orders, it has nothing to learn from.
• This is the "cold start" problem. New stores, new products, and new visitors all lack history.
• It can feel random early, before it has enough signal.
• If it counts only what other customers bought, it can pair a product with a bestseller that does not belong with it (see the Baymard finding below).

How much data does automatic pairing need? Nobody publishes a measured threshold. Shopify says its purchase-history recommendations need previous sales, and it falls back to product descriptions and related collections without them. Growth Suite's rule of thumb is to start with manual picks below roughly 100 orders. Treat any number, ours included, as a starting point, and look at what the algorithm actually shows before you trust it.

Does Shopify have frequently bought together built in?

Partly. Shopify's free Search & Discovery app controls the recommendation blocks that compatible themes show on product pages. According to Shopify's Help Center article on customizing product recommendations:

• Complementary products ("products often bought in addition to a selected product") are the closest thing to an FBT box. You choose up to 10 per product, and you can edit them in bulk with metafields. Your theme needs a complementary products block, and only compatible themes have one.
• Related products (similar items, potential substitutes) are generated automatically for every product. Shopify says its built-in strategies use products commonly purchased together (which needs previous sales), similar product descriptions (English storefronts only), and related collections as a fallback. The article describes automatic generation only for related products.
• Hybrid is built in. For related products you can show only your own picks, or your picks ahead of the generated ones. Shopify recommends keeping the generated ones on, so a slot is never empty when a hand-picked product sells out.
• Shopify's own advice for complementary picks: choose products that are a good add-on and lower than or equivalent in cost to the original.

The article describes recommendations only. It does not describe bundle pricing, an add-all button, or reporting on what a pairing earns, which is the gap dedicated apps fill.

So which one actually converts better?

We could not find an independent head-to-head test of manual vs AI Frequently Bought Together. This is what exists:

• A vendor average, with limits. Growth Suite, a Shopify app vendor, says in its February 2026 cart upsell guide that AI cart suggestions average a 3.8 percent add-to-cart rate, versus 1.56 percent for manual picks. The guide gives no sample size or method, and it measures suggestions in the cart, not an FBT box on a product page. It points in one direction. It does not prove it.
• One store's report of what FBT earned. In the Oxify Cart Drawer & Upsells analytics for one store (screenshot in our product page upsell guide), frequently bought together offers brought in $8,294.15 of $26,763.27 in product page upsell revenue, behind cross-sells at $12,847.32. It shows FBT can earn. It does not compare manual with automatic pairing.
• A usability warning about automatic pairing. Baymard Institute's 2021 research found that 52 percent of the desktop sites it benchmarked show cart cross-sells that are completely irrelevant or based only on what other customers bought. In its tests, even one questionable recommendation made users lose faith in all of the suggestions and ignore them. Baymard tested cart cross-sells, not FBT boxes, but the lesson carries over: an algorithm that only counts co-purchases needs exclusion rules and hand-pinned pairs behind it.

Our view, not a measured result: the offer matters at least as much as who picks the products. A pairing at full price gives the shopper little reason to take both, so a small bundle discount (5 to 10 percent is a starting range we suggest, sized against your margin) is the first thing we would test, ahead of switching to AI. We have no data that separates the discount effect from the pairing method.

So the sensible order of priority is:

1. Give shoppers a reason to take both items, such as a bundle price.
2. Place the box in high-intent spots: product page (under Add to Cart), cart, and post-purchase.
3. Then pick manual, automatic or hybrid based on your order history.

Show 2 or 3 products that truly fit, never a wall of random items.

Manual vs AI: side by side

Which tool should you use, and what does it cost?

Most stores that outgrow the free option reach for an app, because a dedicated tool adds bundle pricing, a one-click add-all button, and reporting. Listing data below was checked in October 2026. Plans and prices change, so re-check before you install.

Both app listings price by monthly store orders, not by store size. For more apps and full breakdowns, see our best Frequently Bought Together apps and best upsell apps comparisons, plus our guide on how to choose a bundle app.

High-ticket stores (like electronics):

• If you sell expensive items, pick a tool that prices by order volume, not revenue. A store doing 150 high-value orders pays the same tier as a store doing 150 cheap ones, even though its revenue is far higher.
• For considered purchases, relevance and trust matter most. Pair automatic suggestions with a few hand-picked accessories, warranties, or add-ons, and keep every suggestion genuinely useful.

How to set up FBT on Shopify (step by step)

You need no code. Plan on about ten minutes for a first version.

1. Check your order count. Under roughly 100? Start manual. Past that, automatic pairing has something to learn from.
2. Pick an app. Shopify's free Search & Discovery lets you choose complementary products by hand. For bundle discounts, a one-click add-all button, and reporting, use a dedicated app.
3. Turn on the app embed. Online Store, then Themes, then Customize, then App embeds.
4. Set your pairings. Choose 2 or 3 complementary products per bestseller, by hand or with AI.
5. Add a bundle discount. Start at 5 to 10 percent and check it against your margin.
6. Place the box well. Product page under Add to Cart, plus the cart and the post-purchase or thank-you page.
7. Test on mobile. Most shoppers are on phones, so check it looks clean there.
8. Watch the numbers (see below). Cut dead pairings, keep winners.

Not sure how upsell, cross-sell, and bundle differ? Our product page upsell guide breaks it down.

How to switch from manual cross-sell lists to AI

Already have manual cross-sell lists and want to move to AI? Do it without losing your best pairings:

1. Export what you have. Note your current manual pairings, especially for your top sellers.
2. Install an app with both modes (manual and AI), so you can run them side by side.
3. Turn on AI for the catalog and let it learn from your order history.
4. Hand-pin your proven winners. Keep the manual pairings that already convert, and let AI fill the rest.
5. Compare for two weeks. Track add-to-cart rate and attributed revenue. Keep what wins, drop what does not.

This hybrid path means you never lose a high-converting pairing while the AI catches the ones you missed. Our guide on adding upsells to your cart drawer walks through the setup.

Benchmarks: how to know it is working

Track four numbers for your FBT box:

• Impressions: how often it shows.
• Click-through rate: how often shoppers click a pairing.
• Add-to-cart rate: how often a click becomes an add.
• Attributed revenue: the actual money it brings in.

No published target fits every store. The only add-to-cart figures we found are the Growth Suite cart averages above (1.56 and 3.8 percent), from a vendor with no method published, so use your own before-and-after numbers instead. If clicks are high but add-to-cart is low, the pairing or its price may be off. Fix the pairing or the placement.

Common mistakes to avoid

• No reason to take both. A pairing at full price gives shoppers little incentive to add the second item.
• Too many items. Stick to 2 or 3. More just confuses people.
• Wrong placement. Burying the box where no one scrolls.
• Set and forget. Pairings and offers need a refresh now and then.
• Pricier add-ons. Keep suggested items the same price or cheaper than the main product, so the "yes" feels easy. Shopify's Help Center gives the same advice for complementary products.

Where Oxify fits

Oxify Cart Drawer & Upsells (our app) runs its FBT block on the product page and in the cart drawer, and the same app runs one-click post-purchase offers. You can set pairings by hand per product or collection, or use automatic recommendations built from your order history, with a manual override per product, exclusions for products you never want offered, and fallback rules so a new product is never left with an empty block. Bundle pricing is applied through Shopify's discount engine, and the analytics report revenue per offer. The Frequently Bought Together page has the full feature list, and the cart drawer page has current pricing.

Who should skip it: if you only want a product-page FBT box, a dedicated app such as Frequently Bought Together CBB is simpler. Oxify has no free plan (it starts at $9.99 a month with a 14-day free trial on every plan), and its 40 App Store reviews are far fewer than the 1,207 for Frequently Bought Together CBB.

Want volume discounts as well? Oxify Quantity Breaks (our app) adds "buy more, pay less" pricing on top of your bundles. (See our best quantity breaks apps guide.) Together the two cover both "buy these together" and "buy more of this."

Try Oxify Cart Drawer & Upsells free for 14 days and see what your own store's pairings earn.

The bottom line

Start manual. Once you are past roughly 100 orders with a catalog too big to pair by hand, add automatic pairing and keep hand-picked pairs for your top sellers. The only published comparison is a vendor's cart figure (3.8 vs 1.56 percent add-to-cart), so treat the gap as a hint, not a result. Whichever method you use, give shoppers a reason to take both items, place the box in high-intent spots, and judge every pairing by what it earns on your own store.

Related Guides

• Best Shopify cross-sell apps, compared by how they pick products

Frequently Asked Questions