Ecommerce Merchandising: Assign Each Surface a Job

Ecommerce merchandising is the decision of which products to feature, bury, or rotate on a discovery surface, using conversion, stock, and exposure, and when to reverse a placement that does not earn the slot. That slot decision stays here. Assortment planning does not own it, and separating Shopify controls does not replace it.

Photos and grid consistency belong in visual merchandising for ecommerce. Reorder quantity and the collection-wide ranking math belong in assortment planning. This page is the slot decision: which product takes a visible place, which one moves down, which one is on a dated rotation, and how you undo a placement that loses. On Shopify, a collection sort, a Search & Discovery product boost, and a homepage section are separate controls you use after that decision is made. Saving one does not change the others.

Choose feature, bury, or rotate

Start with one product and one surface. Give it one of three jobs. Feature means move it into a slot shoppers actually reach. Bury means move it out of that slot. Rotate means give it the slot for a dated window, then put the previous product back unless the test earns a longer stay.

Use three signals, in this order. Do not pick the product because the photo looks stronger.

Conversion against peers in the same collection. Calculate it with the product conversion rate method, then compare that product with the others already in the collection. A storewide average will promote a weak product that only looks fine next to the whole catalog.
Stock cover for the window you are about to open. Days sales in inventory tells you whether on-hand stock can support the extra demand a visible slot can create. If cover is thin, do not feature it. Bury it or leave it where it is until a reorder lands.
Exposure you can actually name. You know a product is first in a collection because you moved it there. You know it is in a homepage section or a search boost because you saved that control. Product sessions do not tell you which of those slots the visit came from.

A product earns a feature when it converts at least as well as the other products in that collection, has stock that covers the placement window, and is not already sitting in the slot you are about to give it. A product earns a bury when it converts worse than those peers, or when it already occupies a prime slot and is not turning that exposure into orders. A product earns a rotation when it is a launch, a seasonal piece, or an under-exposed converter you want to test, and you can name the date you will take it down.

Read the three signals before you edit

Conversion: compare inside the collection

Open the collection you are about to change and rank its products by product conversion rate, not by revenue alone. Revenue favors products that already sit at the top. Conversion tells you which product turns a visit into an order once a shopper reaches it.

Feature a product that converts with its peers or better and is buried below the first screen of that collection. Bury a product that sits in the first rows and converts worse than the products under it. Do not feature a product with almost no sessions yet. Wait until it has enough visits in that collection to compare, or put it on a short rotation instead of a standing feature.

Stock: cover the window, then stop

A featured product will get more demand than the same product in row twelve. Before you move it, check whether available stock covers the days you plan to leave it up. Use days sales in inventory for that check. Do not rebuild the formula here.

Cover runs past the end date you are about to set, and the product converts with its peers: feature or rotate.
Cover runs out before that end date: do not feature. Leave it, or bury it if it is already in a prime slot and you would rather sell a product you can restock.
The product is unavailable: bury it on every surface you control. Search already sinks sold-out products. A theme section will keep showing the last product you saved until you change it.

If the bury is really a reorder problem, hand that SKU to the reorder pass in assortment planning. Do not use a homepage section to hide a stockout.

Exposure: name the slot, do not infer it

Write down where the product sits today before you move it. Collection position is a fact you can see in the collection editor. A homepage section is a fact you can see in the theme editor. A search boost is a fact you can see in Search & Discovery. A rise in product sessions is not a fact about any of those slots. Sessions can come from an ad, a direct link, or another collection.

Treat a prime slot as occupied. If row one already holds a product that converts as well as the candidate and has stock, do not bump it for taste. Rotate the candidate into a dated test, or feature it on a different surface that is empty.

Put the decision on the surface that can carry it

Once you have already chosen feature, bury, or rotate, apply that choice on one control. The control does not choose the product for you. A change in collection order does not create a search boost. A homepage section does not reorder a collection. Shopify documents collection order, including manual sort, drag, and Move, in Editing collection details and layout.

Decision	Collection grid	Search	Homepage or campaign section
Feature	Default sort set to Manually, then drag or Move toward the beginning	Product boost only for terms that product truly matches, and only if it is available	Add it to the section you already use for featured products
Bury	Move it toward the end, or out of the first screen	Delete or narrow the boost. Do not leave sold-out products boosted into unrelated terms	Remove it from the section the same day it should stop
Rotate	Move it up, write the end date, and write the product that goes back	Create the boost, record a deadline, and manually delete the boost on that date unless you extend the deadline in writing	Swap the section product, record a deadline, and manually revert the section on that date

Collection order is the standing slot

From Shopify admin, go to Products, then Collections, and open the collection. In Collection items, set Default sort to Manually. Drag the product, or select it and use Move to send it to the beginning, the end, or a position number, then save. New collections default to Most relevant. Best selling sorts by all-time orders. Neither of those is a feature decision, and neither updates search.

Use manual sort for the feature or bury you just decided. A standing collection rule can stay without an end date. A rotation in the grid still needs one, because Best selling will not put the old product back for you.

Search boosts answer one query

A product boost assigns search terms so that product ranks higher when a shopper uses those terms. Shopify's guide to modifying search with Search & Discovery says the boost applies only if the product is available, sold-out products are listed at the end of search results, and a boost is not applied when the query contains search syntax. You can add up to 10 search terms. Boosting a large number of products can push other relevant products lower.

Boost a product only after the feature decision is already yes, and only for terms a shopper would use for that product. A boost does not drag it to the top of a collection. If the product is hard to find before you boost it, check the title, product type, and tags first. A boost lists it above the usual results for the terms you assign. It does not fix a catalog shoppers cannot describe.

From Shopify admin, go to Apps, then Search & Discovery.
Click Search, then Product boosts, then Create product boost.
Select the product you have already decided to feature or rotate.
Add only the terms that match that product. Save.
Search the storefront for that term and for one unrelated term. The product should appear for the first and not lead the second.

Homepage sections and recommendations

Homepage and campaign layouts are theme sections. Saving a section does not change a collection's Default sort, and it does not create a product boost. Update the section yourself when the product becomes unavailable or the end date arrives. Search settings that sink out-of-stock products do not edit the theme for you.

Recommendation blocks are a fourth surface. Follow Shopify product recommendations for setup and the Behavior reports. Do not treat a recommendation rule as a substitute for a collection bury. A new product should also clear the product launch checklist before it takes a homepage section or a broad set of search terms.

Measure the placement, then reverse it

A placement is a test with a written end, even when you hope it becomes a standing feature. Treat one to two weeks as editorial guidance and an operating heuristic for this workflow, not a published benchmark. The basis is operational: that span is long enough for a homepage section, a new boost, or a grid rotation to produce orders you can compare with the peers you displaced, and short enough to reverse a losing slot at the next weekly review. Shorten it when the product is seasonal or stock is thin. Extend it only by writing a new end date after you have read the result.

Log five fields before you save the edit: surface, product, what moved out of that slot, start date, and end date. Hold other edits still. A discount, a collection reorder, and a new boost in the same week will all move orders, and the log will not tell you which one did it.

On the end date, compare the placed product with the same peers you used at the start. Look at orders and conversion for that product, and at whether stock still covers another window. Do not treat a rise in sessions as proof the slot worked. Sessions are a rough proxy for demand. They are not slot exposure.

It converts with its peers or better, and stock still covers the next window: keep the feature, and write a new end date if the surface was a rotation or a boost.
Conversion is worse than the peers you displaced, or orders did not justify the slot: reverse it the same day. Move the previous product back in the collection, delete or narrow the boost, and put the old product back in the theme section.
Stock will not cover another window: reverse the feature even if conversion looked fine. A slot you cannot fulfill is a bury, not a win.
You changed two surfaces at once: you cannot call a winner. Reverse the extra change, keep one surface, and run the window again.

Reversal is the same control you used to place the product. Collection order goes back with another manual Move. A boost comes off in Search & Discovery. A theme section changes only in the theme editor. Leaving a losing boost in place because the collection move was reverted does not finish the test.

Run the review once a week

Open the log, not the theme, first. The weekly pass is still the same decision: feature, bury, or rotate, using conversion, stock, and the slot you named. For each open placement, check the end date, stock cover, and conversion against the peers in that collection. Then make one of three moves: keep and redate, reverse, or leave it alone until the end date. Do not add a new feature in the same pass as a reversal on the same surface.

List every manual collection position you are treating as a feature, every product boost, and every homepage section that names a specific product.
Mark any placement with no end date that is not a standing collection rule. Give it an end date this week or take it down.
Bury anything unavailable, and anything whose cover will not last to its end date.
Reverse anything whose conversion is worse than the product it displaced.
Pick at most one new feature or rotation for the coming week, on one surface, with the peer comparison and the stock check already written down.

Mistakes that waste the slot

Featuring the product with the best photo, or the highest revenue, without comparing conversion inside that collection.
Featuring a product whose stock will not last the window, then calling the stockout a traffic win.
Reading product sessions as visits from a known row, boost, or homepage slot.
Boosting a hero into unrelated queries, or boosting a large set of products for a broad term. Shopify's own guidance is to keep that list small and specific.
Changing collection order, a boost, and a homepage section in one afternoon, then judging the week as one test.
Letting a rotation run past its end date because nobody wrote down what to put back.
Handing the feature, bury, or rotate call to assortment planning, or stopping once you have named which Shopify control is which.

This week, pick one collection. Name one product to feature, one to bury, and one rotation with an end date. Change only the surface that matches that decision, and put the reversal on the calendar before you save.