Recommendations for every product page.
A sleeping pad and a stove beside the tent, other trail shoes beside the trail shoe, and what shoppers bought with it. Findlane builds each product's lists from your catalog, so they work on the first day, before a single order or click.
Featherbed UL Pad
Flint Canister Stove
Owlet 20 Down Bag
Spark 200 Headlamp
Glade 2 Tent
Waypoint 2 Tent
Tamarack 3 Tent
Kinglet 1 TentThree lists for each product.
Goes with and Similar are built ahead of time from your titles, categories, descriptions and prices. Bought together comes from your orders. A product page reads every list in one request.
Rill Squeeze Water FilterGoes with: what to use it with.
Products to use with it or for the same occasion: a sleeping pad, a stove and a headlamp for a tent. A model decides which categories go together, and how sure it is of each pick, from product text alone.
- No orders, clicks or tracking needed, so a new store or a new product has lists from the start.
- A price limit keeps a $249 cooler off a $54 stove's page: picks cost at most 3× the product by default, or 2×, 5×, or any price.
- The category map shows which categories the model says go together. Switch a pair off, and that category's lists are rebuilt.
Talus Ridge
Bracken 2 Trail
Linnet Trail
Longhaul UltraSimilar: what to choose instead.
Alternatives for the shopper who likes the trail shoe but not its price, its fit or its color. Other sizes and colors of the same product come after the alternatives, or not at all.
- Filters apply to the picks as they do to a search, so
inStock: truekeeps sold-out products off the page. - A change of price or stock rebuilds nothing. Filters and the price limit read each product's current record when a page asks.
- Up to 20 products a list, and 8 by default.
Trailbed Midweight Merino Hiking Sock
Scree Running Soft Flask
Swiftline 5 Running Vest
Boulderfield Low Trail GaitersBought together, from your orders.
Send orders from your server, and import past ones, and each product gets the products that were in the same orders, the most shared orders first. Shopify stores choose Use my orders in the app instead.
- A product shows in another's list once at least 2 orders had both, so a list can't give away one customer's order.
- A list shows your few best sellers twice at most before its other picks, so pages don't all show the same products.
- Orders from the last year count, without orders of over 20 products, such as wholesale ones. New orders show in about an hour.
Cornice Mid, new this week
Loam Crew Sock
Splitpace Cap
Pocketline Belt
Pacer 500A full row, even with few orders.
A product with few orders still gets a full row: fill tops Bought together up from Goes with, in the same response. The heading says what most of the row is, Frequently bought together or Goes well with.
- In tests on stores with 4,000 to 22,000 orders a year, most product views had a Bought together list. A store with a few hundred orders a year had lists for few products, which is what
fillis for. - A list renders nothing while the product has none, so the same component is safe on every product page.
- A product's lists come in one request, which the edge caches like a search.
Ridgepot Titanium Cook Set$64
Longreach Titanium Spork$14.95
Cinder Double-Wall Titanium Mug$38
Holdfast 45 Rotomolded Hard CoolerSee every list before shoppers do.
The Recommendations page shows how many products have lists, and each product's lists, with the picks that were left out and why. Try it first: every workspace can build lists for its 20 most popular products free, shown in the dashboard only.
- Lists are built in the background, the most popular products first: by a
popularityfield, or by your orders and views. - New and changed products get lists in the next build, which runs every few minutes after writes. Deleted products leave every list.
- Builds run apart from search, so searches and record writes never wait for them.
- The clicks on each list show there too, once your pages send them.
Built from your catalog.
No tracking script, and no wait for data. These are the fields the lists read, and Findlane finds them in most catalogs.
categoryA title and a category
What each product is, and the field that groups products, which you can choose. A description helps too, most when titles are short.
priceA price
For the price limit, so Goes with picks stay near what the product costs. A price change rebuilds no list.
popularityA popularity number
Such as units sold in the last 30 days, so your best sellers get lists first. Without one, orders and then views decide.
On a product page, in two lines.
React's component draws each list and sends its clicks. The browser and server SDKs return the lists as data, for markup of your own.
import { Recommendations } from "@findlane/browser/react"export function ProductPage({ product }: { product: Product }) { return ( <> <ProductDetails product={product} /> {/* Bought together, topped up from Goes with */} <Recommendations objectID={product.id} list="together" fill="goesWith" /> <Recommendations objectID={product.id} list="similar" /> </> )}
const { ready, goesWith, similar, together } = await client.recommendations("hollowpine-larch-2")if (!ready) return // its lists aren't built yet
const lists = await products.recommendations("hollowpine-larch-2", { limit: 6, filters: { inStock: true },})lists.goesWith[0].score // 0.94: how sure the model islists.together[0].count // 14 orders in the last year had both
Good to know.
- An add-on to a paid plan, from $6 a month for 4,000 products, billed apart from the plan, monthly or yearly.
- It covers your most popular products, up to the number you choose. A growing catalog never costs more by itself.
- Requests aren't counted as searches. Browser requests are rate-limited like browser searches.
- On Shopify, add it in the Findlane app and place the Recommendations block on product pages.
- Bought together covers every product with orders, including those beyond the products your add-on covers.
- Goes with and Similar read only product data. Product text is processed by models on Cloudflare, where Findlane already runs.
Questions, answered.
Do I need orders or tracking before they work?
No. Goes with and Similar are built from your catalog alone: titles, categories, descriptions and prices. They work the day you turn them on. Bought together adds lists from your orders once you send them.
How is it priced?
By the products it covers, in units of 1,000, billed apart from your plan: $1.50 a month per 1,000 up to 10,000 products, and less per 1,000 beyond. At least 4,000 products, which is $6 a month. A year costs ten months. It needs a paid plan.
What if my catalog is bigger than what I pay for?
Recommendations cover your most popular products, up to the number you choose, so the products shoppers see most get lists first. A growing catalog never costs more by itself: the dashboard offers to cover the rest.
Does it work on Shopify?
Yes. Add recommendations in the Findlane app, which puts them on your plan's Shopify charge, and place the Recommendations block on product pages. For Bought together, choose Use my orders.
Will it slow down my product pages?
A product page gets all its lists in one request, which the edge caches like a search. Lists are built ahead of time, apart from search, so searches and record writes never wait for them.
What do shoppers' orders reveal?
Nothing about who bought what. Goes with and Similar use only product data. Bought together uses which products were in the same orders, with no customer details, and a product shows only once at least 2 orders had both.
More features.
Merchandising
Pin, boost and hide products on category pages and searches, and run campaigns that start and end by themselves.
Read moreSearch analytics
What shoppers search, what finds nothing and why, where they click, and what search sells.
Read moreTypo tolerance
Misspelled, half-typed and run-together searches still find the product, with a correction or related results when nothing matches.
Read moreEvery feature
Search, filters, merchandising, recommendations, analytics, and the rest.
See them allSee your top products' lists free.
Every workspace can build lists for its 20 most popular products and preview them in the dashboard. Turn recommendations on for the rest when they look right.