Ecommerce store on subdomain - danger of keyword cannibalization?

[Alces]

Hi all,

Scenario: Ecommerce website selling a food product has their store on a subdomain (store.website.com). A GOOD chunk of the URLs - primarily parameters - are blocked in Robots.txt. When I search for the products, the main domain ranks almost exclusively, while the store only ranks on deeper SERPs (several pages deep).

In the end, only one variation of the product is listed on the main domain (ex: Original Flavor 1oz 24 count), while the store itself obviously has all of them (most of which are blocked by Robots.txt).

Can anyone shed a little bit of insight into best practices here? The platform for the store is Shopify if that helps. My suggestion at this point is to recommend they all crawling in the subdomain Robots.txt and canonicalize the parameter pages.

As for keywords, my main concern is cannibalization, or rather forcing visitors to take extra steps to get to the store on the subdomain because hardly any of the subdomain pages rank. In a perfect world, they'd have everything on their main domain and no silly subdomain.

Thanks!

[effectdigital]

That sounds like a hell of a mess. Instead of tying your name to one proposed implementation and saying "yes, this IS the way" - I'd get the complexity of the issue across to the client / boss

I'd then present your idea and say "I want to test this, but if results suffer we will need to revert the changes". I think that with such a complex architectural nightmare (on a HORRIBLE platform like Shopify, which is just awful for SEO) - it would be extremely foolish to charge off into the night without making the risks clear

The best practice is really to not have built such a terrible site to begin with. In making things better, there may be growing pains. There may be NO options which would result in 100% growth and 0% losses

My recommendation would be to continue blocking Google's access to the original, default product variations (as those are already happily ranking on the main site. Don't fix what ain't broken). I might allow Google to crawl the sub-variations which are inaccessible from the main site. I might alter the main site's UX to include links to the sub-variants on the 'shop.' subdomain

In the end though, it's a very tangled web they have spun

[Alces]

Yeah, I'm trying to figure out the best way to present to them all the pertinent information regarding how terrible Shopify is. The way they use Collections then block any sort of parameters in their unalterable Robots.txt file is insane.

[ClaytonJ]

One reason we got out of shopify.  Gets complicated quickly.   There was a brilliant WBF on subdomains about 2 months ago - by the british dude from distilled who pops up from time to time.  Will try and find it if get time, but would check that out as a starting point.

[ClaytonJ]

In summary - firstly echo effectdigital on Shopify.  It is an interesting platform sold very well by Shopify zealots - but we have had to bend too many times to Shopify platform limitations to believe it is the right answer for most.  It is awesome if your a bikini start-up with no CRM or ERP - however the moment it comes to a decent integration - it often gets ugly quickly.

On to your query - the shortened version to the answer is no-one knows.  Why? because the algorithm treats subdomains differently for different sites. https://moz.com/blog/interview-searchlove There is a good piece on subdomains v subfolders in this WBF.  In summary a good discussion on subdomains.

The click through to the subdomain should be a normal step, ie so assuming on the subdomain your landing on the relevant contextual page within the funnel to transact.   That is normal for some back ends. You are correct ideally in my view all on the root domain.

Overall if the subdomain pages are critical and you want to rank, then need to treat subdomain for SEO as a separate site.  However, if the subdomain is just the end part of the sales funnel.. then may not need to rank..

Hope that is helpful.

Regards

[Alces]

This was incredibly helpful. Right now their funnel starts on the store (adding product to cart), but there's definitely a benefit to it starting on the main domain to better track how the channels perform and overall user behavior.

[effectdigital]

I posted a bit of a Reddit rant here under my personal SEO alias of "studiumcirclus":

https://www.reddit.com/r/SEO/comments/bgqelg/changing_web_host_will_it_affect_google_rank/em1m0cg/?context=3

(click "View Entire Discussion")

Mainly these things vex me about the platform:

"In basic terms, Shopify is limited by its vision. They want to make sites easy to design for the average-joe, which means they have to spend most of their platform dev time on the back-end of the system and not the front-end of the sites which it produces

_ If they're always bogged down making extra tick-boxes to change things in the back-end, how can they be keeping up with cutting edge SEO? With WordPress you have a much larger dev community making add-ons, many of them completely free and still very effective. Because everyone is on WP, when new Google features, directives or initiatives come out they are quickly embraced (putting all sites on WP one step ahead)_

_ With smaller dev communities, platforms like Shopify or Magento lag behind. Why do people always expect that 'average' will rank 'well'? Ahead of the curve ranks well, average ranks averagely_

_ Also Shopify has some nasty Page-Speed issues which they won't acknowledge and they just argue about instead of fixing things. It's just not good for SEO_"

Other "Shopify is bad" evidence:

https://moz.com/community/q/main-menu-duplication#reply_391855 - just contains some of my thoughts on why Shopify isn't that good

https://moz.com/community/q/site-crawl-status-code-430 - a relatively recent problem someone had with their Shopify site, scroll down to see my reply

https://moz.com/community/q/duplicate-content-in-shopify-subsequent-pages-in-collections - someone else having tech issues with their Shopify site. While my answer was probably right, they probably couldn't implement the fixes