Huge drop in Search Queries overnight - any ideas why?

[gavinharriss]

I've noticed a sizable dip in traffic with [my site] so headed over to Google Webmaster Tools to see if Google had encountered any issues with the site.

Was quite surprised to discover that the "Search Queries" stats dropped considerably after Oct 29th. Any ideas why Google might suddenly have decided not to favor the website so much?

There's no messages about penalties in Google Webmaster Tools. No links have been bought for the site, though many links do exist out in the wild that tend to be highly relevant. No black hat SEO techniques have ever been employed.

The site has existed in it's present form (content within tabs to the left) for a few years now without Google issues. I was originally worried that Google may view this content as hidden black hat SEO text, but it doesn't seem to have been an issue so far and I'd hope Google is a bit cleverer with it's search ranking signals due to the recent popularity of single-page-sites.

Any ideas what may have caused the abrupt drop in Search queries?

Thanks in advance for any insights!

[edited by: brotherhood_of_LAN at 11:49 pm (utc) on Nov 10, 2013]
[edit reason] See stickymail about personal links, thanks! [/edit]

[netmeg]

I was originally worried that Google may view this content as hidden black hat SEO text

Why would you think this?

You maybe have been hit by one of the rolling Panda algorithm updates.

[gavinharriss]

Looks like the site url has been removed from my original post so I'll describe the site content to give better context:

The site is a full page interactive map with no visible text based content initially. Content is available via initially closed tabs that pop-out from the left of the page when clicked by the user. I used this approach to maximize the map real-estate on screen which is the real core content useful to the user.

I was initially worried a few years ago when the site went live that Google may have interpreted this as hidden injected black hat SEO content. But it never seemed to be an issue previously.

I guess without being allowed to post a link to my site it's hard to get relevant feedback or suggestions :( The site is a bit different to the usual text based content websites.

I also posted a link to a screenshot (with no information that would identify the website) that showed the huge sudden drop in Search Queries stats, but this was also removed from the post. Guess I have a bit of learning about what is acceptable to post on this forum.

[brotherhood of LAN]

gavinharriss, see the message at the top of the page for you explaining the link policy, it'll give you an idea as to why we avoid links where possible on here... including protecting you from other people linking to your site to highlight problems they want others to see, it does happen.

Some questions
- Was the text spidered by Google, could you search for strings on the page and the page would be found?
- Is the text content already in the source code when the page is fetched, or is it fetched via AJAX? You mention visible text based content initially but that doesn't clarify whether you mean visually not there or not in the source code at all.

[gavinharriss]

Fair enough, didn't think about abuse of links :)

I've just checked and Google does in fact seem to be indexing the hidden content (did a few quick searches for text in Google search), and all content is available on initial page load (no AJAX content loading).

[brotherhood of LAN]

That's good to know, at least they are seeing that you have content on the page. Do you get any traffic from sources other than Google search?

[gavinharriss]

Here's some numbers just to give an idea of traffic source breakdown percentages:

Organic Search = 25,281
Direct = 9,917
Referral = 6,676
Social = 1,347

[brotherhood of LAN]

Good to see traffic from various sources. Worth trying the 'fetch as googlebot' to ensure that there's no spidering issues.

Also check your backlinks, perhaps you lost some of your quality backlinks there were underpinning the authority of your site, or perhaps a competitor has threw a load of garbage links at your site. Either way it's definitely worth checking.

Shooting in the dark here as I'm not an SEO per-se, but it's worth mentioning these kinds of checks to rule out anything that'd be a big red flag.

[gavinharriss]

There doesn't appear to be any spidering issues.

Backlinks - there's only the odd garbage link. Also compared to my only real competitor I seem to have 1878 mostly highly relevant backlinks compared to their 475.

The competitor is now outperforming me in search results. My site previously appeared on page 1 for certain search terms and now it's page 3, where-as the competitor is still page 1.

The competitors site hasn't changed much in the last 2 years.

I did notice that results in search results that were outperforming me were using a - for url spaces, while I was using %20. I've since changed this on my own site to see if -'s perform better.

Thanks for the input.

[gavinharriss]

A stab in the dark...

It wouldn't be anything to do with the company the site might keep being in a shared hosting environment, potentially sharing an IP address with sites Google has taken a dislike too?

[brotherhood of LAN]

It's possible but highly unlikely that your shared hosting would be affecting your rankings, Google's familiar with the fact that lots of unrelated sites can sit on the same IP and I imagine the IP would have to host a lot of unsavoury sites to even be considered a factor.

I did notice that results in search results that were outperforming me were using a - for url spaces, while I was using %20. I've since changed this on my own site to see if -'s perform better.

Did you do this before or after your ranking drop? Either way it's not a great idea to change URLs like that. What happens to requests to the old %20 URLs?

[bluntforce]

"search results that were outperforming me were using a - for url spaces, while I was using %20. I've since changed this on my own site to see if -'s perform better"

Science at it's best. Change a character across a site, then try to differentiate the character change effects vs. page change effects.

One controlled change at a time. Limit the change to a small environment.

Or, just make random changes and believe you can interpret the effects of your changes.

[gavinharriss]

I changed to -'s after the rankings dropped (not before), all the old %20 urls still work as the ids in the url is what the website uses, the rest of the text is just for SEO. Canonical urls all now point at the new -'s versions.

@bluntforce - Glad you like my style ;) Good points though - taken on board.

[bluntforce]

Why would any search engine value a site higher due to characters used in an url? A really crappy URL that provides exactly what a user wants is what a search engine should accept and embrace.

You can change whatever you want, stupid changes for the sake of change don't address base issues.

[gavinharriss]

Maybe it has some ideals about url best-practices - who knows eh. But I guess this would be a very minor signal compared to the huge drop I just witnessed. Very curious what might have happened, or what new signal is at play.

[brotherhood of LAN]

If both your - and %20 style URLs are returning 200 status codes, that's an issue in itself that commonly causes problems for site owners, as Google will now see lots more URLs within your domain but not any more new content - therefore duplicate content. Not sure what the preferred method is here but my suggestion would simply be to go back to using your old URL structure, as that's where the majority of your link juice is and where Google originally knew your pages were.

If this is a common trait in your URLs it may have been a major factor in your ranking drops, consider putting a canonical tag on the pages just to be explicit about what your intended URLs are.

[gavinharriss]

Okay, I'll revert back as the urls have been changed less that 24 hours. Does seem that I jumped the gun a little on that one! Cheers.

Though I would hope the religious use of canonical url meta tags on all pages should prevent Google being confused?

[brotherhood of LAN]

> canonical url meta tags on all pages should prevent Google being confused?

Yes. Case sensitivity in the URL path or query, extra query variables, www as a subdomain, trailing slashes... there are loads of ways for Google to get confused, particularly when you return a 200 response to more than one variation. The canonical tag takes the mystery out of it.