TL;DR for busy store owners
Google often indexes filter combinations instead of your categories and products, because filters get more internal links than anything else on your site.
Some filters should rank. The job is deciding which ones, instead of letting Google decide.
Run every filter through six questions. Pass all six: it becomes a landing page. Fail one: keep it out of Google.
Never use noindex and a robots.txt block on the same URL at the same time.
You can check whether you have this problem in 10 minutes in Search Console. The steps are below.
Here’s an uncomfortable truth about your online store: Google has probably already decided which pages represent your business. And there’s a good chance it didn’t pick your categories.
This summer I audited 197 online stores. In 52 of them, indexation was the single biggest problem, and it looked almost identical every time. Google had indexed hundreds, sometimes thousands, of filter combinations. Size plus colour. Colour plus price range. Price range plus “sort by newest”. Meanwhile the category and product pages that actually make money were missing from the index or buried behind all that noise.
Nobody on those teams chose this. It happened one filter click at a time.
The conversation with the owner usually goes the same way. They open Search Console expecting to see their best categories at the top. Instead they find URLs like ?colour=black&sort=price_asc. Then they check the category they spent weeks improving, and it says “Crawled, currently not indexed”. That’s the moment this stops being a technical problem and becomes a money problem.
The 30 second test
Open Google and search for this:
site:yourstore.com inurl:colour
Replace “colour” with any filter parameter your store uses: size, brand, price, sort, filter. If you see pages of results that look like filter combinations, keep reading.
The site: operator gives you a rough sample, not an exact count. That’s fine. You’re not measuring yet, you’re checking whether the pattern exists.
What most store owners believe (and why it’s half wrong)
Belief one: “Filter pages in Google are a bug. Noindex all of them.”
This sounds safe. It isn’t.
Some filter pages match real searches better than your category ever will. Someone searching “women’s waterproof hiking boots” doesn’t want your Hiking Boots category with 340 products. They want exactly the filtered result. If that filter page has a clean URL and a proper title, it can be one of the best landing pages on your site.
To be fair to the other side: ranking filters is not the problem. Some of your filters should rank.
Belief two: “Google will figure it out. The canonical tag handles the rest.”
This one is more expensive.
A canonical tag is a hint, not a command. Google ignores it when it thinks two pages are different enough. Filter pages usually are: different products, different counts, different content.
What these beliefs cost you
Googlebot spends its attention on URLs that don’t sell.
Your internal linking power is spread across thousands of near duplicates.
Your categories and products compete with their own filtered versions, and often lose.
So the real question isn’t whether filters should rank. It’s who decides which ones. Right now, in most stores, the answer is Google.
What’s really happening under the hood
Filters win for a simple reason: they get more links than anything else on your site.
Every category page links to every filter option. Every page of pagination repeats those links. Now multiply. One category with 6 colours, 8 sizes, 4 sort orders and 5 price ranges can generate 960 URL combinations. That’s before multi-select, where a shopper picks two colours at once and the number explodes again.
Your category, by comparison, gets a handful of links from the main menu. Your products sit three or four clicks deep behind pagination.
Google follows links, and it crawls and indexes what it finds most often. That’s how filter combinations end up representing your store.
The trap most teams fall into
When teams finally notice, the usual reaction is to add noindex to filter pages and block them in robots.txt, all at once. It feels thorough. It backfires.
If a URL is blocked in robots.txt, Google can’t crawl it. That means it never sees the noindex tag. Those URLs can stay in the index for months with the label “Indexed, though blocked by robots.txt”.
By now you’re probably thinking: fine, so what do I actually do?
The honest bridge
If “noindex everything” is wrong, and “let Google figure it out” is how you got here, what’s left is a decision process. One that you control.
I’ve run the same process across 197 stores this summer. The pattern is consistent enough that I can give it to you as a checklist and a worked example.
To be transparent: fixing this is part of what I do at Lobit, and there’s an offer at the end. But everything below works whether or not you ever talk to me.
The “deserves to rank” test: six questions
Run every filter, or filter combination, through these six questions. Six yeses means it’s a landing page. Any no means it’s a filter that stays out of Google.
Is there real search demand for exactly this combination? Check your keyword tool and your own Search Console queries. “Waterproof hiking boots women” has demand. “Hiking boots, blue, size 38, sorted by price” doesn’t.
Will the page stay full? If the combination shows three products today and zero next month, it becomes a thin or empty page. Set a minimum product count and stick to it.
Does it have a clean, static URL?
/hiking-boots/women/waterproof/is a page.?attr_12=4&sort=price_ascis the state of a widget.Does it have its own title, H1 and short intro, written for that search? If the title still says “Hiking Boots | Store Name”, Google has no reason to treat it as a separate page.
Does it have a self referencing canonical and a place in your XML sitemap? This is how you tell Google the page is intentional.
Is it linked from the parent category as a normal link? Not only through the filter widget, but as a visible link that a shopper (and Googlebot) can follow.
Filters that pass get treated like categories. Filters that fail get handled in the cleanup steps below.
A worked example: one category, from 960 URLs to 11 pages
Let’s take a store selling outdoor gear. The category is Hiking Boots, with these filters: gender, waterproof, brand, colour, size, price range and sort order. Here’s how the six questions play out.
Gender (women, men): clear search demand, stable inventory. Becomes a landing page. That’s 2 pages.
Waterproof: strong demand (”waterproof hiking boots”). Landing page. That’s 1 page.
Gender plus waterproof: “women’s waterproof hiking boots” is a real search. Landing page. That’s 2 pages.
Brand: people search “Salomon hiking boots”. Only brands with enough stock pass question two. In this store, 6 brands pass. That’s 6 pages.
Colour: almost nobody searches “blue hiking boots”. Stays out. (In fashion, colour often passes. The test decides, not the filter type.)
Size, price range, sort order: no search demand, constantly changing results. Stay out.
Any combination of three or more filters: stays out.
Result: 11 landing pages Google should know about, instead of 960 combinations it was guessing between. Each of the 11 gets a clean URL, its own title and intro, a sitemap entry and a link from the Hiking Boots category.
The kill list: filters that almost never deserve to rank
Sort orders (price, newest, popularity, rating)
View options and products per page
Price sliders and price ranges
Multi-select within the same attribute (red plus blue)
Combinations of three or more filters
“In stock only” and “on sale” toggles
Session, tracking and internal search parameters
If your store creates crawlable URLs for any of these, that’s where your crawl budget is leaking.
How to clean it up, in the right order
Order matters. Doing step five before step two is how stores lose traffic they didn’t know they had.
Step 1: Make the list. Run the six questions and write down which filter pages pass. Check Search Console first: any filter URL that already brings clicks has passed question one in practice. Don’t kill it without a replacement.
Step 2: Upgrade the winners first. Filters that passed get clean URLs, unique titles, H1s and intros, sitemap entries and links from their parent category. If an old parameter URL for the same filter already has traffic, redirect it to the new clean URL.
Step 3: Stop creating new junk. For filters on the kill list, stop handing Google crawlable links. Google’s own documentation on managing crawling of faceted navigation URLs recommends preventing crawling of filter URLs you don’t need indexed, for example with robots.txt rules, or by using URL fragments, which Google generally doesn’t crawl. How you implement it depends on your platform and developer (see below).
Step 4: Let the existing junk drop out. Filter URLs that are already indexed get a noindex tag and stay crawlable until they disappear from the index. Only then, if you need to, block them in robots.txt. Never both at once.
Step 5: Strengthen categories and products. Link key categories from the menu and from related categories. Reduce click depth to products. Make sure every product you want found is in the sitemap.
Step 6: Check monthly. Indexation is not a one time fix. New filters, new apps and plugins, and new product lines quietly recreate the problem.
Platform notes: Shopify and WooCommerce
The platform changes how you fix this, not what you fix.
Shopify
Collection filters create parameter URLs, and filter apps often add their own.
Shopify’s default robots.txt already blocks some sort and multi-tag combinations. Don’t assume it covers your filters. Open
yourstore.com/robots.txtand check what’s actually there.To turn a filter into a real landing page, the most reliable route is usually a dedicated collection (for example “Women’s Waterproof Hiking Boots”) with its own URL, title and description. Then link it from the parent collection.
Your filter parameters should stay out of the index.
WooCommerce
Watch for these parameters:
orderby(sorting),filter_plus an attribute name (layered navigation), andmin_priceandmax_price(price filter).Filter plugins often add their own parameters on top.
Filters that pass the test are usually best served as attribute archive pages or dedicated subcategories with written intros, rather than parameter URLs.
Check your SEO plugin’s settings to see which of these URLs it lets into the sitemap.
Custom platforms and other systems
Ask your developer one question: “Which filter combinations create a link in the HTML with a unique URL?”
The answer is usually the size of your problem.
Parameter hygiene: the small things that multiply the mess
Even filter URLs you keep crawlable should follow a few rules:
One order, always.
?colour=red&size=38and?size=38&colour=redare two URLs to Google. Pick one order and enforce it.Lowercase, consistent separators. Mixed case and different separators create duplicates.
Empty results shouldn’t return a normal page. If a combination has zero products, it shouldn’t serve a thin page with a 200 status. Return a 404 or keep it out of the index.
Landing filters need real links. If your filters are buttons that only update the page with JavaScript, Google may never find the landing versions. The 11 winners need normal
<a href>links.
Seven mistakes I see in audits all the time
Noindex plus a robots.txt block on the same URLs. Covered above. Google never sees the noindex.
Canonicalising every filter to the parent category, while also wanting them to rank. Pick one. A page can’t be both “ignore me” and “rank me”.
Canonicalising paginated pages to page one. Page 2, 3 and 4 are how Google reaches the products on them. Tell Google they’re duplicates of page one and those products get harder to find.
Noindexing pagination. Same result, different route.
Using nofollow on filter links as the main fix. Google treats nofollow as a hint, not a crawl control.
Removing filter URLs that bring traffic, without checking Search Console first. This is how a cleanup turns into a traffic drop.
Fixing it once and never checking again. The next plugin update brings it all back.
Check your own store in 10 minutes with Search Console
Pages report. Open the indexed pages and look at the example URLs. Then open “Crawled, currently not indexed” and “Discovered, currently not indexed”. If your categories and products live there while filter URLs are indexed, you have the pattern.
Performance report. Filter pages by URL containing your filter parameter. Compare their impressions and clicks with your main categories.
URL Inspection. Run your five most important categories through it. If any say “not indexed”, or show a Google selected canonical that isn’t the page itself, start there.
Crawl stats (Settings, then Crawl stats). Look at the example URLs Googlebot requested recently. If most of them contain parameters, you can see where your crawl budget goes.
Larger stores can go one level deeper: crawl the site with a tool like Screaming Frog and count URLs by parameter, or analyse server log files to see exactly which URLs Googlebot requests.
How to measure whether the fix worked
Track three numbers before you start and every month after.
1. Index coverage by page type. Submit separate XML sitemaps for categories, products and filter landing pages. Search Console then shows indexing status per sitemap. You’ll see immediately whether your categories are in or out, instead of guessing from one big total.
2. Where Googlebot spends its time. In Crawl stats, the share of requests going to parameter URLs should fall over time.
3. Clicks to categories and your chosen landing pages. In the Performance report, these should grow while clicks to random filter URLs fade out.
The index shifts first, traffic follows later.
Why this also decides your AI visibility
AI assistants that search the web before answering, like ChatGPT and Perplexity, lean heavily on search indexes to find pages worth citing. If your index is full of random filter combinations, that’s what they find when they look at your store. Your best category page never enters the conversation.
In my audits I track three numbers for this:
AI Visibility Rate: how often a store appears in AI answers for its core queries.
Citation Rate: how often it’s cited as a source.
First-pick Rate: how often it’s the first recommendation.
Stores with a messy index almost always score low on all three.
If you use Bing Webmaster Tools, its AI Performance report now shows Citation Share. That’s a free way to start seeing where you stand.
What changes when you fix it
Googlebot spends its time on pages that sell, so updated categories and new products get picked up faster.
Your categories stop competing with their own filtered copies, and the signals that were spread across hundreds of URLs land in one place.
Product pages get indexed, which opens up the long tail of specific product searches your competitors are catching.
Chosen filter pages capture high intent searches your categories were never going to win.
Your Search Console data becomes readable, because you’re no longer analysing 4,000 URLs to understand 40.
You give AI assistants something worth citing, instead of a sort order.
Questions store owners ask me about this
Why not just noindex all filters and be done with it?
Because some of your filters are your best landing pages for specific searches. Noindexing them all removes the problem and the opportunity together. Check which filter URLs already bring clicks before you touch anything.
Can’t my developer fix this in an afternoon?
The technical part is often small. The decision part is the hard one: which filters pass, which URLs currently drive traffic, and in what order to change things so nothing breaks. If you have the time and the data, you can absolutely do it yourself with the checklist above.
How long until I see results?
The index shifts first, traffic follows later. How fast depends on how often Google crawls your store. Anyone who promises you a date without looking at your crawl data is guessing.
What if some filter pages already bring traffic?
Keep them. That’s evidence they pass question one. Upgrade them into proper landing pages, and redirect the old parameter version to the clean URL.
How many filter landing pages should one category have?
As many as pass all six questions, and no more. In the hiking boots example that was 11. In a large fashion category it could be 40. In a niche B2B category it might be 2.
Does this apply to small stores too?
Yes, and sometimes more. A store with 200 products and 30,000 filter URLs has a worse ratio than a big store with the same problem.
Want me to check your store?
You’ll get:
which filter URLs Google has indexed
which of your key categories and products are missing or buried
the first three filters I’d turn into landing pages, and the ones I’d take out of the index
It comes as a short written note within 5 working days. If your index is clean and there’s nothing worth fixing, I’ll tell you that too, in one line. You lose nothing but the time it takes to send an email.
Remember the math from earlier: every week Googlebot spends crawling hundreds of filter combinations is a week it’s not recrawling the category you just improved.
Not ready yet? Forward this to your developer or whoever manages your store. The six question test and the seven mistakes are worth 20 minutes of their time. And if you want to save this, bookmark it. It’s written to be a checklist, not a one time read.
Neven
P.S. Filters ranking in Google isn’t the problem. Google choosing which ones, while your categories and products sit outside the index, is. Reply with your store URL and I’ll show you which side of that line you’re on.

