On a 3,000-SKU Shopify store, the default search returned nothing for 14.2% of queries. After 30 days of synonym rules, typo tolerance and redirects, Boost AI brought that down to 3.6%, Searchanise to 4.1%, and SoBooster to 4.8%. Search conversion rose from 2.9% to between 3.8% and 4.2%. The app mattered less than the cleanup work: about 80% of the gain came from fixing the top 200 failed queries, which any of the three apps can do.
A zero-result search is one of the clearest signals in ecommerce. A shopper told you exactly what they wanted, and your store said "nothing found". Many of them leave right there.
On small stores, it rarely matters. With 200 products, shoppers browse. On large catalogs, search becomes the main way people shop, and a high zero-result rate quietly costs sales every day.
We worked with a home and garden store carrying about 3,000 products. Their search report showed one in seven queries ending with no results. Here is how we fixed it, and how three search apps compared doing the same job.
- Why searches fail on large catalogs
- The four-step fix we used
- How Searchanise, SoBooster and Boost AI compared
- What to check every month
Why do searches return zero results?
We exported 500 failed queries from the store's search report and grouped them by cause.
Synonyms made up 38%. Shoppers typed "hose pipe" while the products said "garden hose". They typed "planter" when the catalog said "pot". The product existed, but the words did not match.
Typos made up 24%. "Lawnmover", "secateur", "fertiliser" spelled the British way on a US store. Shopify's default search handles some of this but not all.
Products not stocked made up 19%. People searched for things the store did not sell. You cannot fix these with a synonym, but you can guide shoppers to something close.
SKU and part numbers made up 11%. Trade customers searched model numbers that were not indexed.
The remaining 8% were mixed: vague words, very long phrases, or test searches.
Baymard Institute's research on ecommerce search query types shows the same pattern across many stores: shoppers use many query styles, and most sites support only a few of them well.
What is the four-step fix?
Export, group, fix, check. It works with any search app that supports synonyms and redirects.
Export. Pull the top failed queries from your search report. Shopify's own analytics show search terms, and search apps add a dedicated zero-result report. Sort by frequency and start with the top 200.
Group. Put each query into one of the causes above. This takes an hour or two for 200 queries, and it tells you where to spend effort.
Fix. Add synonym rules for the synonym group. Turn on or tighten typo tolerance for the typo group. For products you do not stock, create a redirect to the closest collection or a helpful page. Add SKU and model number fields to the search index.
Check. Rerun every fixed query yourself. Then watch the zero-result rate over the next two weeks.
On this store, the first round took about six hours. The zero-result rate dropped by more than half before we touched any advanced features.
How did the three apps compare?
We ran each app for 30 days with the same synonym list, then let each app's own features add what they could.
Boost AI: 3.6% zero-result rate, 4.2% search conversion. Its AI-assisted suggestions caught some synonyms we had missed, and its merchandising rules let us pin best sellers for vague terms like "outdoor".
Searchanise: 4.1% zero-result rate, 4.0% search conversion. Strong typo handling and a clear zero-result report. Bulk synonym import from a spreadsheet worked well.
SoBooster: 4.8% zero-result rate, 3.8% search conversion. Fast and simple, with good filter integration. Fewer automatic suggestions, so it depended more on our manual list.
The store's default search, for reference, sat at 14.2% and 2.9%.
The gaps between apps were around one point. The gap between any app with cleanup and no cleanup was ten points.
Which features matter most?
Four features did the heavy lifting.
Synonym rules with bulk import. You will add hundreds. A spreadsheet upload saves hours.
Typo tolerance you can tune. Too loose, and "rake" matches "cake stand". Too strict, and typos still fail. All three apps let you adjust this.
Search redirects. When someone searches "returns" or "delivery", send them to the policy page, not a product grid. When they search a product you do not sell, send them to the closest collection.
A zero-result report. You need to see what still fails each week.
Pricing for all three starts around $19 to $29 a month for a catalog this size and rises with products and search volume. Check current plans on each app's listing.
Boost Commerce, the company behind Boost AI, also offers filter features that overlap with Shopify's Search & Discovery app. We covered those in our Boost AI Search & Filter review.
How should you write synonym rules?
Synonyms are the biggest lever, so it pays to write them carefully.
Use two-way synonyms for true equivalents. "Hose" and "hose pipe" mean the same thing. Searching either should show both.
Use one-way synonyms for broader terms. "Planter" should show pots, but "pot" should not always show every planter box. One-way rules prevent results from getting noisy.
Do not overreach. Mapping "chair" to "furniture" sends shoppers to a broad page when they wanted chairs. Keep synonyms specific.
Include regional spellings. "Colour" and "color", "fertiliser" and "fertilizer", "aluminium" and "aluminum". If you sell across countries, these matter.
Add brand nicknames. Shoppers often type a shortened brand or a common misspelling of it.
On this store, the final list had about 340 synonym rules. Roughly 60 came from the apps' suggestions, and the rest from the failed-query export.
How do you handle SKU and part number searches?
Trade buyers and repeat customers often search by code, not by name. On this store, 11% of failed queries were model numbers like "GH-2040" or supplier part codes.
Shopify's default search indexes product titles, descriptions, tags and some variant fields, but SKU matching can be patchy, especially with dashes and spaces. A shopper typing "GH2040" may not find "GH-2040".
All three apps let you add SKU, barcode and metafields to the search index. Turn that on first. Then add a rule that ignores dashes and spaces in codes, if the app supports it, or add the common variants as synonyms.
We also added old SKUs as tags on products that replaced discontinued ones. Customers reordering an old part now land on its successor instead of a dead end.
After these changes, SKU searches with zero results fell from 55 a week to 6.
What did search changes do to revenue?
Search users are a small share of visitors, but they buy more.
On this store, about 18% of sessions used search. Those sessions produced 31% of revenue before the cleanup, because shoppers who search usually know what they want.
After the fix, search conversion rose from 2.9% to about 4% across the three app periods. With roughly 9,000 search sessions a month and an average order of $62, that change was worth around $6,000 a month in extra revenue.
Not all of that came from fewer zero-result pages. Better ranking of results and redirects for service queries helped too. But the zero-result fixes were the first and largest step, and they took the least time.
What should a no-results page show?
Even after cleanup, some searches will fail. The page they land on decides whether the shopper stays.
Baymard's guidance on no-results pages recommends offering a way forward rather than a dead end. We followed that with four elements:
- A clear message that repeats the search term
- Spelling suggestions when the app has one
- Popular categories as large tiles
- Three to six best sellers
Search exits from the no-results page dropped by about a fifth after the change. The best sellers block did most of the work, since many shoppers clicked one of those products and kept browsing from there.
Does search work with Shopify's built-in tools?
Shopify's Search & Discovery app is free and supports synonyms, filters and product boosts. For many stores under 1,000 products, it is enough.
On this 3,000-SKU store, it brought the zero-result rate from 14.2% to about 7% with the same synonym list. The paid apps went further because of stronger typo handling, SKU indexing, redirects and reporting.
Start with the free app. Move to a paid one if your failed-query report still shows many typos and SKU searches after cleanup.
For a side-by-side look at the three paid apps on features and price, see our Boost AI vs Searchanise vs SoBooster comparison.
How do you keep the zero-result rate low?
Search is never finished. New products, seasons and trends bring new queries.
We set a monthly routine on the store:
- Export the top 50 new failed queries
- Add synonyms or redirects for each
- Check five random popular searches by hand
- Note the zero-result rate in a simple sheet
Each round takes about 45 minutes. Six months after the first cleanup, the store's rate stayed under 4%.
When new product lines launch, add their common names and nicknames to the synonym list on day one. That avoids a spike in failed searches during the launch week, when traffic is highest.
Large catalogs also benefit from good collection structure, so that search and browsing support each other. Our Warehouse vs Empire theme test covers menus and collection trees for stores with thousands of products.
Questions merchants ask us about this
What is a good zero-result search rate?
Under 5% is a solid target for large catalogs. Many stores start between 10% and 20% before any cleanup.
Do I need a paid search app to fix zero-result searches?
Not always. Shopify's free Search & Discovery app supports synonyms. Paid apps add stronger typo tolerance, redirects, SKU search and reports.
How many synonym rules does a store need?
It depends on the catalog. Our 3,000-SKU store ended with about 340 rules after two rounds of cleanup.
Should I redirect searches for products I do not sell?
Yes, to the closest collection or a helpful page. It keeps shoppers moving instead of leaving on a dead end.