Punished by Google for updating links in old posts

[mboydnv]

In a separate post I will address in detail how I did positive things for my site only to see my traffic constantly erode.

In the meantime, a very strange thing happened the other night. What little traffic I had vanished to nothing. Here’s what happened.

A few months ago, I moved my site to https:// - So on Dec 15 around 7pm PST, I went to my google plus account and changed the urls in the posts from a year ago to now. I simply updated the links from http ://www. to https:// - 20 minutes later i go to view keyword activity in Statcounter, and for 20 minutes I’m seeing all kinds of action (maybe 40 different hits).

I'm seeing:
www.google.com(Keywords Unavailable) #1
www.google.ie(Keywords Unavailable) #4
www.google.ca(Keywords Unavailable) #2

and so on. The IP was all Google. And various Google locations throughout the world, not just California.

What I find odd is statcounter including the bot as a unique visitor. If it was a bot. The IP did say Google, but it was done through various browsers and resolutions.

Shortly after, My Google traffic eroded down to 4 hits in maybe 36 hours. I’ve been in business 10 years with this website.

And yesterday I paid $20 for stumbleupon ad, and got a 100 views and not one like. Can’t seem to get anything to get shared no matter what we write, social traffic does not work.

Anywho, thought I’d share what happened. Was it a bot? Was it a group of human editors punishing my site for updating old links in Google Plus. It doesn’t make sense. I’m cleaning up old links, wouldn’t they want me to do that?

[edited by: goodroi at 7:49 pm (utc) on Dec 17, 2015]
[edit reason] Delinked http [/edit]

[dipper]

" ... thing happened the other night" - what date?

I ask because the first step should be to compare your *events* to Google Panda rollout (observed) dates.

It's worthwhile doing this because your assumption on these things being the problem might be incorrect.

[not2easy]

Not addressing all the questins, but regarding what you are seeing. If you switched from http ://www. to https:// you would also need to add the https: version to the Google Search Console (old GWT) and replace the old version of statcounter with one for https: URLs, in order to get accurate information on your site from either Google or Statcounter. Also, is it possible you overlooked changing the URL in your WP "Settings" file? New URL = New Site.

[lucy24]

The IP did say Google, but it was done through various browsers and resolutions.

There are several googloid functions, mostly but not exclusively having to do with mobiles, that come through with google IPs. If you've got access to your request headers, Google is pretty good about including the "X-Forwarded-For" header so you can see where the human visitor really comes from. Normally it will turn out to match whichever national google was in the referer.

[mboydnv]

dipper: Dec 15 around 7pm PST
not2easy: All went well with the change to https.

[dipper]

It might be panda - I suggest you look into your traffic history vs google algorithm history.

This is a helpful article on it - [searchengineland.com...]

Also, keep in mind that the recent panda rollout was started a number of months ago and was noted making impacts around October 1st and afaik is still ongoing - with sites seeing new impacts almost every 2 weeks.

[netmeg]

Google doesn't generally move that quickly on a manual penalty even for people who deserve one. I suspect it was whatever update (Panda or something else) that supposedly rolled out earlier this week.

[tangor]

I know we all want to know "cause and effect" scenarios as that would make is so much easier to do things and experiment, but the reality is the web, these days, is so fluid, and for the g watchers, so involved with penguin this panda that and still overtones of florida and all that other happy g stuff....

Sometimes it is only as simple as present a new face to the web (move from http to https, for example) that "restarts" the SEs in their indexing, which means, behind the scenes, other algos are running to see if this "makes sense".

What are your Bing metrics? You can't figure g out if you don't have another example which can either correlate, or disprove what you see in g. Duck Duck Go and Yandex also come to mind. If those also show the same fail on your site then you did something very wrong, otherwise, it's just g being g.

Do know this, g will never tell you. That's one of those "live with it" kind of things.

[mboydnv]

Hi Tangor,

I think my problem may have been panda, thin content. My problem was woprdpress images not attaching to posts. Because of deleted posts, creating rogue images. etc

I did a Google Search for my site as site:mysite.com and discovered many urls being indexed as: mydomain.com/?attachment_id=1202

I removed those from the engine, reattched images and deleted images not being used. crossing fingers for next update. However GWT now reporting many ?attachment_id=XXXX as 404. Why can't they leave me alone?

[lucy24]

However GWT now reporting many ?attachment_id=XXXX as 404.

Huh. When you said "removed" I though you meant you'd changed a setting in the parameters area, telling them to ignore this parameter. If you're sure the settings are correct, this is a good example of a wmt message that can safely be ignored. (Remember, it's mainly informational. Not all messages mean you did something wrong.)

Question for anyone who happens to know: If you tell google in wmt/gsc to pick some random value of a given parameter, do they still report 404 for any other values they happen to try?

[not2easy]

If you tell google in wmt/gsc to pick some random value of a given parameter, do they still report 404 for any other values they happen to try?

Yes they do for awhile, if you mean other parameters that are in that URL. I've never checked exactly how long, but then it tapers off and goes away. My best estimate is "under 60 days".

You know how Gbot never forgets an URL it has seen, well they may keep looking for it, but if it was set to "ignore" in parameters then they won't keep bothering with 404 notifications. If it was simply "removed" they won't quit. They may slow down for a while and later it will all be back.