An SEO professional was concerned about their latest site migration.
At the beginning of December, they had migrated a CMS site that averaged 50,000 sessions per day. It did not have AMP, and had many technical and SEO problems.
Their migration resulted in a full AMP implementation that had 301 redirects that directed to the final pages, which were all in full AMP format.
Even though they did this, and the professional believes their implementation is 100 percent correct, they lost a lot of traffic as a result.
John explained that what he usually does with these types of questions is, he consults Google’s internal tools for data that would look similar to what you would find in Google Search Console.
Then, he compares before/after periods of changes to see if he can find anything that stands out, in terms of things that may not have been ported over to the new site properly.
He finds that sometimes, the impressions match and sometimes they don’t match.
The one thing he found, regarding this particular site, is that it ranked for many terms that weren’t really all that relevant to the site’s main topic.
For example, if you have questions like the following:
- How to log in to WhatsApp,
- How to install WhatsApp,
- How to download WhatsApp,
But the site itself is not all about WhatsApp. From John’s perspective, the site is not really something that matches the query.
So what happened was that the site had a lot of traffic for these queries, but at some point, it’s likely a bug on Google’s side that the site was ranking for these queries.
Because the site wasn’t the official WhatsApp site.
He went on to stress that they don’t just show one official site for these types of queries.
However, if you were ranking for these types of queries, but you’re not exactly a site that caters to the overall product under that banner, then you can’t really rank for those queries in a legitimate way.
This happens at approximately the 33:18 mark in the video.
John Mueller Hangout Transcript
John (Submitted Question) 33:18
Then, at the beginning of December, we migrated a CMS site. The site averaged 50,000 sessions per day. It didn’t have AMP, and it had a lot of performance and SEO problems. The migration consisted of making respective 301 redirects going to a full AMP format. And it goes on and basically says, since the migration, the website has basically lost all traffic, or a lot of traffic.
John (Answer) 33:46
So I took a quick look at this before. Usually what I do with these kinds of questions is go to our internal tools, and where I see something similar to what you would see in Search Console. And I try to look at a period before the change, and after a change, in particular, with regards to kind of like the overall visibility that the site gets, so the impressions.
And sometimes, that kind of matches what when a site actually makes technical changes, and sometimes it doesn’t really match. And one of the things that I noticed when looking at this site in particular, is it used to rank for a lot of terms where I would consider it not to be that relevant. So for example, what I saw was that it was ranking for things like how to log into WhatsApp, or how to install WhatsApp, how to download WhatsApp, those kinds of queries and the website itself is not actually WhatsApp.
So from my point of view, that’s something where essentially, this website got a lot of traffic for queries, where I would say, it could be considered almost like a bug on our side that this happened. And obviously, we show a number of different sites for these kinds of queries. And it’s not that we just show that one official site for that. But it’s still something where I would say, if you’re not WhatsApp, and you’re kind of like suddenly no longer ranking for queries, like how to log into WhatsApp, then from that point of view is like, it’s not really a bug on our side that this has changed.
And sometimes these kinds of queries can drive a lot of impressions and with them, probably a lot of traffic as well. So if you just look at the overall traffic to a website, it can look like there was a significant change. But if you look at the actual kind of relevant traffic for a website, assuming I think this is like something like a news type website, if you look at kind of like the relevant traffic that is relevant for this kind of website, then that looks like it’s actually pretty stable.
And it’s, it’s sometimes tricky to figure out, like, which part of the traffic that you’re seeing, or the impressions that you’re seeing is actually relevant to the website. And it’s hard to also have a high level view of the website while trying to ignore those less relevant things. So, for example, we saw this with our own search documentation, where we work together with the SEO folks to try to de-emphasize specific queries.
In particular, I think it was like the query, I don’t know, Google video or something like that, where lots of people were searching for Google video. And we had a page on video structured data for Google, which technically is kind of like matches those keywords. But practically, it didn’t really make sense for those users. So when we looked at the overall impressions to our site, we saw this fairly high number.
But actually, that wasn’t really useful for us as a metric because it contained this big segment of traffic for things that are kind of irrelevant for our site. And that, I think, is a bit tricky. And means that when you recognize this kind of situation, it means that you can’t just look at the overall traffic to a website and say, “Oh, this is a metric that should remain stable or go up.” You should try to figure out what is the portion of relevant traffic there? And how can I recognize that and track that over time?
So from my point of view, looking at this specific site, I didn’t see anything technically incorrect, with regards to this migration. It’s essentially just our search systems, trying to figure out like, what should we really show for queries like this? And maybe it’s not that particular site? And probably it’s not really a bug that we would not show that. And my feeling is that the timing here was almost more accidental with regards to the technical changes that were made. And it’s something that would have happened anyway, over time.
