'%20clip-path='url(%23canvas-clip)'/%3e%3cg%20clip-path='url(%23canvas-clip)'%3e%3cg%20transform='rotate(0,%20256,%20256)'%3e%3cg%20transform='translate(87%2087)%20scale(14.083333333333334%2014.083333333333334)%20translate(0%200)'%3e%3cpath%20fill='none'%20stroke='%23FFFFFF'%20stroke-linecap='round'%20stroke-linejoin='round'%20stroke-width='2'%20d='m18.064%2010.877l-4.89-7.26c-.42-.625-1.287-.803-1.936-.397a1.4%201.4%200%200%200-.41.397l-4.893%207.26C4.24%2013.715%204.9%2017.318%207.502%2019.423a7.13%207.13%200%200%200%204.168%201.572M15%2018a3%203%200%201%200%206%200a3%203%200%201%200-6%200m5.2%202.2L22%2022'/%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
ParticleSearch Query Tools: Ranking, Synonyms, Redirects, and Safe Publishing
Ranking rules, synonyms, and redirects solve different search problems. Choosing one because it is available can improve a screenshot while making the wider search experience less predictable.
ParticleSearch Query tools are designed for a narrower workflow: start with a specific query, inspect the evidence, select the smallest responsible intervention, compare the draft with live behavior, then publish and follow up.
Query-tool behavior checked July 28, 2026. This guide covers the current merchant-visible Ranking, Synonyms, Redirects, Live, Draft, Compare, and publishing workflow. Exact controls can differ by product release and store state.
Chapter 1 · Choose the intervention
Diagnose the failure before adding a rule
The question is not “Which tool can change this query?” It is “What is wrong, and what is the smallest change that addresses that problem without rewriting unrelated behavior?”
Observed problem
The right product exists but appears too low for one important query
Consider
Ranking
Why
Change order or visibility without changing what the query means.
Verify first
Verify the product is eligible and relevant before pinning, boosting, demoting, or hiding.
Observed problem
Buyers use a true substitute term for the same product concept
Consider
Synonym
Why
Connect equivalent merchant and shopper language.
Verify first
Both terms should be interchangeable in this catalog context, not merely related.
Observed problem
The query is clearly asking for a destination or policy page
Consider
Redirect
Why
Send navigation intent to the page that directly satisfies it.
Verify first
Use a narrow phrase and confirm it should bypass product results.
Observed problem
The product is missing, incorrect, unavailable, or lacks trusted source data
Consider
Catalog repair
Why
A query rule cannot make a bad merchandise record trustworthy.
Verify first
Repair Shopify source data, eligibility, variant identity, or inventory first.
Observed problem
The query is rare, ambiguous, or already produces a defensible result set
Consider
No change
Why
Rules add maintenance and interaction risk.
Verify first
Record the review and wait for stronger evidence.
Rules are not catalog repair. If the intended product is missing, unavailable, incorrectly structured, or carries the wrong variant data, repair the Shopify source record and confirm the index before adding query-specific behavior.
Ranking asks
Which relevant products should appear first?
The meaning and destination stay the same. Only product order or visibility changes for the query.
Synonyms ask
Which words mean the same product intent here?
The vocabulary expands. The terms should remain interchangeable across the products they retrieve.
Redirects ask
Should search become navigation for this exact phrase?
The product result path is replaced by one product, collection, or page destination.
The decision in practice
Use the rule that matches the failure, then stop
A good rule has a small scope and a clear reason for existing. The examples below are deliberately ordinary because the quality of a query tool is tested in the decisions merchants make every week, not only in unusual edge cases.
Query
“red dress”
Problem
The catalog returns relevant dresses, but a current in-stock product is buried below weaker matches.
Responsible answer
Use ranking if the product already belongs in the result set and the reason for its stronger position is current and reviewable.
Avoid
Do not use a synonym. The query language is already correct.
Query
“trainers”
Problem
The store uses “sneakers” in product data while shoppers repeatedly search for “trainers.”
Responsible answer
Use a synonym only after both terms produce a compatible product set in this catalog.
Avoid
Do not connect a broad category, use case, or attribute merely because it is related.
Query
“returns policy”
Problem
The phrase is repeated and the shopper is looking for a stable information page, not a product.
Responsible answer
Use a redirect when the destination is current, specific, and clearly the better answer.
Avoid
Do not redirect ambiguous phrases that can describe a product or collection.
Query
“PM-12V-5A”
Problem
The exact code should identify a power supply, but the result is empty or shows adjacent accessories.
Responsible answer
Repair product and variant eligibility or field coverage before adding a query rule.
Avoid
Do not broaden an exact identifier into a synonym that hides the data problem.
Query
“blue”
Problem
The catalog distinguishes blue, navy, and slate as separate option values.
Responsible answer
Leave the terms separate unless merchant evidence proves that shoppers use them interchangeably.
Avoid
Do not collapse option values to make a result count look healthier.
Chapter 2 · Ranking
Ranking changes order or visibility, not meaning
Use ranking when the result set contains the right merchandise but its order or visibility is not defensible for a specific query. A ranking rule should have a clear owner, reason, query scope, and review condition.
Before changing order, confirm availability, variant identity, product eligibility, and the natural result set. If a product does not belong in the set, boosting another item will not fix the underlying retrieval problem.
Pin
Place a specific eligible product in a fixed leading position for the query.
Best suited to
A product that must lead a narrow, known query for a defensible reason.
Risk
A stale pin can outrank a better in-stock or newly relevant product.
Boost
Increase a product’s preference without promising a fixed position.
Best suited to
A relevant product that should compete more strongly while preserving room for other results.
Risk
Broad or stacked boosts can make the final order hard to explain.
Demote
Reduce a product’s preference while keeping it available to the result set.
Best suited to
A technically matching product that is usually a weaker answer to this query.
Risk
The item can become difficult to find when it is correct for a narrower shopper need.
Hide
Remove a product from this query’s results.
Best suited to
A clearly inappropriate result when the underlying catalog record must remain searchable elsewhere.
Risk
Hiding can conceal a catalog or field problem instead of repairing it.
Why a merchant changes ranking
Default relevance answers a general retrieval question. Merchandising sometimes adds store context the product record alone cannot express, such as a current campaign, a preferred opening product, or a known weak first result. The rule is justified only while that context is current and the product remains relevant.
A relevant product deserves a stronger opening position
A new seasonal collection, a merchant-approved hero product, or the product that best satisfies a narrow query can lead without excluding other relevant choices.
Boundary: The product must already belong in the result set. Ranking should not turn unrelated inventory into a match.
A technically matching result creates a poor first impression
An accessory, replacement part, unavailable item, or visually weak card may appear before the product the query most directly describes.
Boundary: Confirm whether source data, availability policy, or product presentation is the real problem before demoting or hiding it.
The store has a defensible merchandising policy
The team may prefer in-stock merchandise, a current collection, or a product with a stronger fit for a specific campaign query.
Boundary: Keep the policy narrow, named, and reviewable. A permanent boost without an owner becomes invisible search debt.
A query needs stable, predictable ordering
A high-value exact phrase may need one product pinned while the rest of the relevant set continues to rank normally.
Boundary: A fixed position should be rare. It can become stale when inventory, price, catalog content, or the campaign changes.
Natural order
A · B · C · D
One draft rule
Boost C for this query
Draft order
C · A · B · D
The visual comparison should make the changed products and positions obvious. The merchant still verifies that every promoted product belongs in the result set.
Chapter 3 · Synonyms
A synonym connects equivalent language
Synonyms help when shoppers and the catalog use different words for the same concept. They are not a general association tool. “Desk” and “office” may be related, but treating them as equivalent can broaden both terms and damage precise queries.
Check existing query behavior first. Punctuation, case, spacing, stemming, or typo handling may already connect the terms. A redundant synonym increases maintenance without changing the outcome.
Good candidate
The terms can refer to the same merchandise concept in the store.
Usually too broad
One is a product type and the other is a room or collection concept.
Catalog-dependent
They may be distinct option values. Collapsing them can weaken precise color intent.
Test before ruling
Normalization may already handle punctuation. A synonym should not duplicate existing behavior.
Do not use synonyms to force a product campaign. If two terms are not equivalent, use merchandising, ranking, a collection, or content that states the relationship honestly.
The synonym workflow is a product-set comparison
ParticleSearch asks for a source query, equivalent terms, and a product preview because the risk sits in both directions. A synonym can recover a shopper term that the catalog does not use, but it can also make a precise catalog term much broader. The product set has to make sense for every term in the group.
Start with the shopper wording
Enter the observed term and inspect the current live behavior. The source query should come from real catalog language, a repeated refinement, or a merchant-known vocabulary gap.
Add only equivalent terms
Build a small group of words that can stand in for the same product intent in this catalog. Keep related categories, use cases, and attributes separate.
Compare products across every term
The preview should show whether each term reaches a compatible product set. A successful query is not enough if the new equivalent introduces unrelated products.
Publish and repeat the storefront query
After publishing, run the original and equivalent terms in the shopper-facing surface. Confirm the product set, variant handoff, and any protected exact queries.
Chapter 4 · Redirects
A redirect answers navigation intent
Queries such as “returns policy,” “size guide,” or a named campaign may be better served by a page than a product grid. Redirect only when the destination is clearly the better answer. A redirect replaces the normal result path, so its trigger deserves stricter review than a small ranking adjustment.
Intent
The phrase asks for a destination, not merely products that share a word.
Scope
The trigger is narrow enough that an ordinary product search is unlikely to be the better answer.
Destination
The target is current, indexable when appropriate, mobile-safe, and directly useful.
Escape
The shopper can still reach product search if the redirect does not satisfy the task.
Maintenance
An owner will review the rule when campaigns, policies, handles, or information architecture change.
Product intent
“waterproof work boots”
Keep a searchable product result set.
Navigation intent
“work boot size guide”
A verified size-guide destination may be the direct answer.
How ParticleSearch turns the redirect decision into a safe publish
The redirect editor keeps the current product result visible while the merchant chooses a destination. That is intentional. The buyer may appear to be asking for a page, but the live result can reveal valid product intent that should not be removed.
Run the exact shopper query
Inspect the products the query currently returns. A redirect removes that normal result path, so the merchant needs to know what will be replaced.
Choose one destination type
Use a public page URL for policies or content, a product handle for one exact product, or a collection handle for one exact browse destination.
Verify the destination and masked rules
Preview confirms that the exact query will redirect. It also exposes existing ranking or synonym rules that the redirect would mask.
Publish, test, and preserve an escape path
Run the query on desktop and mobile, confirm the destination is current, and make sure shoppers can still reach product search when the page does not finish the job.
Chapter 5 · Compare and publish
Draft review separates an editable idea from live search
ParticleSearch keeps the current result, the proposed behavior, and their differences visible before publication. This does not make a rule correct. It gives the merchant enough context to reject a draft that solves the target query by damaging the result set.
Live now
The currently published behavior and product set provide the control.
Draft preview
The proposed rule shows what would change before shoppers receive it.
Compare
The merchant inspects added, removed, moved, and redirected outcomes.
Publish
Only a reviewed, valid draft becomes live behavior.
Follow up
New evidence is compared after enough time and demand have accumulated.
Publish when
- • The query intent and failure are understood.
- • The intervention matches that failure.
- • Live and draft differences are expected.
- • Important adjacent queries still pass.
- • An owner and follow-up condition are recorded.
Reject or revise when
- • The rule compensates for broken catalog data.
- • A redirect masks valid product intent.
- • A synonym connects related rather than equivalent terms.
- • The target query improves by weakening protected queries.
- • The team cannot explain why the change should exist.
Published rules are an exception inventory
The Published rules page groups live product-ranking rules, synonym groups, and redirects so the team can see which exceptions sit on top of ParticleSearch defaults. Ranking rows show the query and action counts. Synonym rows show the active term group. Redirect rows show the exact query and destination.
Deleting an exception returns that query or wording group to the managed default behavior. This makes removal part of normal maintenance, not a failure. A rule should disappear when its campaign ends, its destination changes, the catalog now solves the problem, or the evidence no longer supports the exception.
Chapter 6 · Protect important queries
Every publish needs a small regression set
Protected queries are the store's known search jobs. Run the relevant subset before a change goes live and again after later catalog, theme, or rule changes. The set should be small enough to maintain and important enough that a failure would matter.
| Group | Examples | Regression to catch |
|---|---|---|
| Exact identity | SKUs, model numbers, barcodes, part numbers | The correct item disappears, loses variant identity, or redirects away from products. |
| Core categories | High-volume product types and collection language | A narrow rule distorts a broad, commercially important result set. |
| Attributes | Color, size, material, compatibility, fit | A synonym erases a distinction that shoppers and variants rely on. |
| Navigation | Shipping, returns, sizing guide, contact | A redirect catches product intent or points to an obsolete destination. |
| Known no-result cases | Unsupported items and deliberately unavailable requests | A rule creates a misleading result only to remove the empty state. |
Start from evidence in the ParticleSearch analytics guide. For wider strategy, use the Shopify search ranking guide and the synonym implementation guide.