Google Updates and SERP Changes - May 2014

[chalkywhite]

---

System: The following 17 messages were cut out of thread at: http://www.webmasterworld.com/google/4659095.htm [webmasterworld.com] by brotherhood_of_lan - 11:06 am on May 1, 2014 (utc -5)

---

I know its Labour Day around the world but my god is quiet today, wouldnt be surprised if an update has rolled out also.

[Mentat]

It was a google update around that time (secret).
It seems that "duplicate content" was one of the problems.

I have another problem since 25 April 2014.

My Structured Data dropped like a stone, no error, nothing.

It's a legitimate site, with 10 ppl working 10 hours/day...

[i.imgur.com...]

[Sand]

I saw a huge uptick in traffic on Tuesday, April 30th - Wednesday, May 1. So far, it has stuck.

This was a site that did have some Panda issues in the past.

[IanTurner]

@andreiut - do not make panic changes to your site, sometimes sites fluctuate wildly in the SERPs for a couple of weeks before the positions stabilise.

[crobb305]

sometimes sites fluctuate wildly in the SERPs for a couple of weeks before the positions stabilise.

IanTurner,

I have seen a very gradual increase in Google traffic over the past few months. I noticed your comment above, and I am curious to know if you have seen this pattern before (screenshot link below). Note the two spikes in February and again in early April. In between those spikes, the trend line seems to be a very gradual upward. In the past week, I've had two very heavy crawls (the highest in 6 months). I honestly don't know what to make of it... and in particular, why my website is spiking in impressions for a day or two before dropping, then repeating the cycle. This is a site that was hit by Penguin 1. It had a manual penalty for unnatural links which was revoked over a year ago (February 2013). No other known penalties exist, I am just waiting for a Penguin refresh.

Here is a screenshot [imgur.com] from GWT.

[ColourOfSpring]

crobb305, same with me - had lots of Google visits in last week or so. A Penguin refresh is due - I don't attribute the extra Google visits to a good or bad result by the way - who knows.

[Martin Ice Web]

@mentant, i think you are right with the "duplicate content" filter. But i think this is not onsite related but similar site ( other websites ) related.
E.G. EAN Numbers a.s.o. are a welcome match key to filter out similar/related pages.

I saw an uptick in traffic yesterday too. Today it seems a bit less but more than the days before. I put some work in my sites to avoid DC.
I guess that google will filter out similar pages more and more. Thats nothing to do with panda or something. i think that it more about efficency. Less cost per single query.
The monthly deep crawlings and then do once a rerun is less expensive than do it all the time.
Now you have only to recrawl the "first" data pool not all the pages.


ecom, germany

[IanTurner]

crobb305 - yes I have seen that pattern before, though your spikes are somewhat higher than the ones I have seen.

I think there is some kind of damping in the algorithm - to prevent rises and falls being too drastic (refer to the early Google monthly updates for the chaos drastic rises/falls caused) and that those spikes are when the undamped results are unleashed on the users for a short period.

I think the trend line is the most important thing - when it is going up you are good, when it is going down - you need to be looking for a cause. If it is flat work on developing the scope of the site.

It would be interesting to know if people have also seen some crevisses (reversed spikes) in their stats and if they are associated with a downward trend line.

[crobb305]

crobb305 - yes I have seen that pattern before, though your spikes are somewhat higher than the ones I have seen.

I think there is some kind of damping in the algorithm - to prevent rises and falls being too drastic (refer to the early Google monthly updates for the chaos drastic rises/falls caused) and that those spikes are when the undamped results are unleashed on the users for a short period.

I think the trend line is the most important thing - when it is going up you are good, when it is going down - you need to be looking for a cause. If it is flat work on developing the scope of the site.

Great points, IanTurner. If there is a dampening factor, the upward trend in between spikes might imply a lessening of the dampening factor with time (perhaps akin to a sandbox?). Based on my traffic logs, I seem to be experiencing another traffic spike today. It will show up in GWT in a day or two.

[IanTurner]

Hmm will try and explain how I see damping happening.

Say there is a significant change in your position in the SERPs for whatever factor which would cause a 50% increase in your traffic overnight - instead of getting all of that in one hit it will slowly release you that traffic at say 10% per month thus bringing you steadily closer to the correct figure.

A similar kind of effect appears to occur with sites suffering from panda - the traffic drop is slow and steady and not one big hit (though I haven't seen any big reverse spikes with panda)

P.S. That traffic graph from your site is close to an ideal position for a site to be in.

[Wilburforce]

instead of getting all of that in one hit it will slowly release you that traffic

What mechanism do you think is involved (how - independently of SERPS - would Google limit your traffic)?

[crobb305]

Say there is a significant change in your position in the SERPs for whatever factor which would cause a 50% increase in your traffic overnight - instead of getting all of that in one hit it will slowly release you that traffic at say 10% per month thus bringing you steadily closer to the correct figure.

IanTurner, do you think the spikes on my traffic graph represent the true potential of my traffic, with the upward trend in between the spikes indicating a move closer to those "true" positions (i.e., without a dampening)? Notice also, that the second spike is larger than the first. If I could extend the chart beyond 90 days, it would reveal the previous spikes were smaller. I think I understand what you're saying... I just hope the trend is a sign of good things to come. I am still only at 5% of my pre-penguin 1.0 traffic. I realize I'll probably never recover completely, but the last Penguin update did bring *some* recovery (up from 0 visitors a day). Hopefully the next Penguin will be friendlier to me. :)

[aristotle]

crobb wrote:
I am just waiting for a Penguin refresh

The last announced Penguin update was Oct. 4, 2013, more than 7 months ago, which is a longer wait than normal. I hope this extra wait time means that they're working on it, trying to fix it.

[andreiut]

@IanTurner - thanks, i disavowed some new strange chinese domains and PR0 domains pointing to my site.

[Shepherd]

last announced Penguin update was Oct. 4, 2013, more than 7 months ago, which is a longer wait than normal. I hope this extra wait time means that they're working on it, trying to fix it.

I think many of us here have a similar hope, I know I do. I got my hopes up for early April, but no such luck. Seems like they have been doing some very brief, intermittent test roll out but nothing solid. I've seen a few sites that lost their sitelinks amid the first penguin have the sitelinks show up briefly and then disappear again over the last few weeks. Not to mention wildly fluctuating rankings.

[crobb305]

I think many of us here have a similar hope, I know I do. I got my hopes up for early April, but no such luck. Seems like they have been doing some very brief, intermittent test roll out but nothing solid. I've seen a few sites that lost their sitelinks amid the first penguin have the sitelinks show up briefly and then disappear again over the last few weeks. Not to mention wildly fluctuating rankings.

I've had very heavy googlebot crawling over the past 7 days. Two very large spikes in crawling over a 5 day period... the largest since October. Could be nothing, but combined with your observations, Shepherd, there may be something in the works.

[IanTurner]

crobb305 - Yes I think those spikes represent the potential of your site and the upward trend is everything moving steadily toward that potential.

Wilburforce - I'm sure the mechanism for the damping is within the SERPs that are delivered, possibly including some dayparting SERPs or delivery to a certain percentage of personalised searches.

[Shepherd]

heavy googlebot crawling over the past 7 days

we had:
small spike 3/13 7,200 pages
large spike 3/24 - 3/29 15,000 pages/day
small spike 4/9 6,900 pages
small spike 4/15 - 4/16 6,800 pages/day
small spike 4/26 5,500 pages

average crawl: 2,000 pages/day

[Sand]

we had:
small spike 3/13 7,200 pages
large spike 3/24 - 3/29 15,000 pages/day
small spike 4/9 6,900 pages
small spike 4/15 - 4/16 6,800 pages/day
small spike 4/26 5,500 pages

average crawl: 2,000 pages/day

How many pages are on your site?

[Shepherd]

How many pages are on your site?

About 100,000 pages.

[defenseup]

I am seeing the same thing.. On May 2nd Google was at it highest on one of my sites.. Then on the 5th my impressions where double what they normally are.. Traffic wasnt any better

[defenseup]

Google on May 2nd crawled 37k pages and on an average day they only crawl an average 4k

[Jez123]

I've seen a few sites that lost their sitelinks amid the first penguin

I lost my sitelinks and they still haven't come back. I wasn't aware that this was a trait of Penguin. I was wondering if my site was still affected by Penguin. I suppose it must be (though most ranking are back but not as strong buut some are still like page 4 etc.)

[crobb305]

this is what my crawls looked like over past week or two (the highest levels of the past 90 days, and the highest since October). These are pages downloaded. My site only has 80 pages, so this is showing 3x that amount.

[i.imgur.com...]

[edited by: aakk9999 at 4:01 pm (utc) on May 7, 2014]
[edit reason] Changed link to go directly to the image [/edit]

[Shepherd]

last 90 days here:

[i.imgur.com...]

[edited by: aakk9999 at 4:00 pm (utc) on May 7, 2014]
[edit reason] Changed link to go directly to the image [/edit]

[Shepherd]

Looking at crawl activity and historical update patterns we should have seen a penguin update early April, I have a feeling that something went wrong with the update...

[jmccormac]

Looking at crawl activity and historical update patterns we should have seen a penguin update early April, I have a feeling that something went wrong with the update...

But would Google ever admit that their twiddlers ever made mistakes?

Regards...jmcc

[Itanium]

I think we actually saw an update, but it f*cked up rankings. Like I said in another thread, I saw huge changes for my website and the competition. Not a big deal, since those changes didn't have any significant effect on the pageviews.

The real problem was the mess with other search results. One time I even got the korean version of a site, after searching on Google.DE. Many other searchterms had similar problems and only returned garbage.

Today I see a lot of movement for various keywords. Let's hope they fixed their mistake.

[EditorialGuy]

I think we actually saw an update, but it f*cked up rankings.

If there was an update (beyond one of the 500 or 600 algorithm tweaks that occur every year), it must have been limited in scope. Then again, most updates are limited in scope, if we're to believe official Google statements like "This update affected 6.8% of English-language queries" or whatever.

For what it's worth, we had an unusually deep crawl by Googlebot on April 17. Normally a deep crawl of our site is followed by an update within five days, but not this time. It's been business as usual for months.

[Martin Ice Web]

@Itanium, I see an uptick in traffic too. Picture search seems up again and arganics serps too.
I noticed that we have a high bounce on daytime and in the night the bounce is significant lower ( i noticed that they change in the nights to a datacenter that has more pages in database ( or less filters))
Sales are only at nights not at daytime. Even google shopping ads donīt end up in sales at days.

I donīt know if it was an update or the many changes we did. But it would be just about 3 month now that we got hit ( 6.february ) and we got released after 3 month ( 6. may ).

ecom, germany

[Shepherd]

Saw another "test" last night, for about an hour at 11:00 pm est (U.S.) traffic from google was up about 80%. Didn't stick, back to normal after about an hour.