Reviews from the web error - where does G pull the info?

[sjr4x4]

One of our clients, in their knowledge panel has 3 sites listed as "reviews from the web", the third of which is our site, which is great.

But it's got the legally wrong name and I'm baffled why Google is displaying it like this? The SERPs show the correct legal entity and I would have thought the data would have been pulled from the same place.

The problem is our client is a member of one trader scheme managed by us, but Google has used the name of a neighbouring scheme. Although this is also managed by us, the client isn't allowed to display that name.

I've checked our meta tags, open graph as well as schema tags with SSDT but can't track down the reason. Is this Google just being Google or am I inadvertently generating the error?

Have attached a screenshot of the SERP result which is named correctly, and the Knowledge Panel which is incorrect (red arrow). Surely both are generated from the same place, which I assume are the schema.org localbusiness and review tags.

Any wisdom before I'm driven to drink?

 

[JoyHawkins]

Hey @sjr4x4

I noticed for several of your search results, Google is appending the title and adding in "Edinburgh Trusted Trader". Usually they do this because that's the name of the website. I know in Wordpress, there is a general setting that you can modify to edit the name of the site. I'm not sure where you would modify that on the platform you're using.

See this article for an explanation.

 

[JoyHawkins]

Mike Blumenthal also pointed out that they show up as a search result when you search in the Edinburgh Trusted Trader section: https://www.trustedtrader.scot/Edinburgh/search.php?bus_type=Carpet+Sun&location=&order=dist

 

[sjr4x4]

Thanks Joy. Platform is homegrown, so no quick easy wordpress button unfortunately

The pointer from Mike is deliberate, idea being vetted traders can show up on across neighbouring schemes, but residents can see which scheme a trader belongs to, ie they are vetted by their local Trading Standards team.

I saw that article about the brand names, but the highlevel brand name is Scotland Trusted Trader not Edinburgh. Basically it's a bunch of official schemes with their own site, code of practice etc, but on the same overall platform.

Google just seems to be adding random site names, unless I've dropped a clanger in a meta tag or structured data somewhere. Just a case of tracing why Google is doing this and where it is pulling the incorrect info from.

Or maybe I need to add a tag somewhere to all the subdomains explaining what the overall sitename is? I don't want to change the og:site_name as it will mess up Facebook shares, which all work fine.

 

[Stefan Somborac]

Steve,

What follows are some ideas and observations, not recommendations! You might want to bounce this off an SEO pro who you trust.

PART 1:
Building from notes from both Joy and Mike:

Mike pointed out that Carpet Sun appears as a search result in the Edinburgh Trusted Trader section. It also shows up if you search in Fife or East Lothian. (But not in Renfrewshire. By design?)

None of those search pages are marked as noindex so Google can index them.

The second result highlighted in Joy's screenshot points to
https://www.trustedtrader.scot/Fife/Carpet-Sun-Limited-0000395.html?reviews=best

Follow that link. The URL you land on suggests that it's a page from the Fife scheme but hovering over the breadcrumbs, it appears this is actually a search result from the Edinburgh scheme. If that's correct, it jives with what Joy mentioned about Google appending page title with the site name.

So, I think that might explain the strange title in Joy's screenshot.

It is commonly recommended that you block your internal search pages from being indexed. Yoast has a post on that. (If the post seems unclear about whether to block those internal search pages with noindex or robots.txt, Google provides the definitive answer: noindex.)

PART 2:
Next, why the strange review in your knowledge panel?

Those three search pages (Edinburgh, Fife, East Lothian) contain review schema; Google can extract a review rating from them. And because those search pages contain a single business listing, marked-up with LocalBusiness and aggregateRating schema, it seems possible that Google might misinterpret those single result search pages as authoritative review pages.

Why then does Google choose the Edinburgh search page over Fife or East Lothian? Well, none of those three search pages contains rel=canonical information. If they're viewed as duplicate content Google picks one as canonical. (Google picks one if you don't specify. See here.)

PART 3:
That leaves me with a question that perhaps someone else on the board can answer:

If you noindex your search pages, Google will still crawl them. If I'm correct about part 2, will Google still use that search results page as a review source even though it's marked noindex?

I ask because I see not indexing a page vs. using a page as a review source as different and independent actions.

You could test by making your search pages noindex.

If the strange Edinburgh review drops from your KP, we've answered my question. (Although this doesn't fix your problem.)

If the strange Edinburgh review remains in your KP, and if I'm right about part 2, removing the schema markup from the search results pages should remove the Edinburgh review from the KP. (Again, not really solving your problem but perhaps uncovering something useful.)

Perhaps the ultimate fix would be to make a canonical declaration on search results pages that return a single listing.

Make sense?

Again, these are just ideas, not recommendations! Making changes of this nature has the potential to go wrong.

Good luck.

 

[sjr4x4]

Thanks Stephan, much appreciated.

FYI, the reason some schemes won't show is down to the search area agreed by the member, so they only show where they can service clients.

You have reinforced the conclusion we're coming to internally, that seems to be the need for canonicals on the search pages, and maybe some no-indexing for internal search, but we'll leave the fixed category pages as indexed. We thought (incorrectly) having canonical tags on each traders page would be enough.

Your breadcrumb spot is fascinating, something we hadn't seen and I didn't pick up from Joys post. We're going to clean up the schema tags while we've got the spanners out, so looks like a busy week next week in the test environment.

Many thanks for the input all, I think I've got what I need to go on, and will update when we've implemented a fix. Looks like it's a bit more complex than I originally anticipated!