Understanding Recent Drops In Google Traffic - Some Questions For You All

[Gorgwatcher]

It seems that a major change is happening in Google since mid of January as reported by many reputable names like whitenight, tedster and others. This thing has been reported by many webmasters and seo's in several other forums other than WebmasterWorld. I am listing down some questions for all of you to please answer. The similarity in answers will decide if this thing is limited to some websites or industry or is this a general thing now. Please answer in yes or no with short explanation.


1. Are your traffic Decline is from usa and UK?
2. Is Traffic Drops only from Google or from other search engines too?
3. Did you notice any decline in GoogleBot Deep Crawling?
4. Is the traffic Drop is sudden or gradual decline?
5. Is your rankings on your major keywords drops?
6. Are your website internal pages caching normally?
7. Decline in Indexing Status Of Your website?


Regards
Wish you all best of luck :)

[1script]

Well Google has now announced that speed has been added to their algo which might explain the recent ranking drops several webmasters here have experienced (active for the last few weeks according to the post over there).

Great catch! Didn't know about that. I stand corrected: it is official now.

What a dumb move! Anyone knows a botnet owner I can hire to slow down a competitor's site?

[1script]

s there anyone here with a really slow site according to Google labs, say average speed > 5 seconds as shown in GWT? If so, have you seen a Google traffic drop over the last few weeks which might be attributable to this?

(I'm talking about the Google labs claimed speed, not about the site's real loading speed)

Yes, most of my sites (both affected by the drop and not) are above 5s load average. Completely not the true load speed but who do you argue it with?

I insert an HTML comment with page render speed into every page, so I can check on it every time. Most of my pages render in 0.02 - 0.10s range. Most pages load approx. 10 pictures, a CSS and JS but it you are visiting the site for the second time, you don't need to load those, just the page itself. This shouldn't add up to 5s, not even close.

So, yes, there is a myriad of questions about how axactly do they measure it. In googletalk it's: "We use a variety of sources to determine the speed of a site relative to other sites."

[vandread]

I have experienced a drop of one main keyword by 10 slots from 3 to 13. My page speed is now about 8 seconds while it was almost 20 seconds a few months ago. Most of that is related to advertisement scripts like Google Adsense that are loaded on my page.

The drop occurred on April 1, not sure if site speed was the reason for the drop or something else. Seems unlikely as all other rankings remained the same. But then, who knows..

[Kelowna]

Google has been showing 2 sets of results so if your site ranks poorly on the new caffeine results and good on the regular results then that would account for the loss of traffic as half the users could be seeing the new as well.

[vandread]

Well it is only that one keyword, everything else remained the same.

[TheMadScientist]

Completely not the true load speed but who do you argue it with?

It's a relative number, so why even worry about what the display number is?

IOW: As long as the calculation is consistent then it does not matter what number they display for people to look at because it's a 'relative to other sites' factor, and as long as the calculation is consistent to arrive at the number it does not really matter what number they display, because the number will be relatively the same for relatively the same perceived (calculated) load speed.

It doesn't matter if they have the number correct to the microsecond (or even the second) as long as they use the same process for arriving at the number across sites, because it's your site's relative load speed compared to other site's relative load speed, so the calculation being consistent is much more important than the number they calculate and/or display.

We use a variety of sources to determine the speed of a site relative to other sites.

Source linked previously by vandread.

[edited by: TheMadScientist at 10:42 pm (utc) on Apr 9, 2010]

[aristotle]

Mad Scientist - I don't sgree about the relativity. Because the method Google uses to measure the "average speed" could favor some types of sites over others. In other words, it could be more inaccurate in some cases than in others, giving erroneous relative values as well as erroneous absolute values.

[tedster]

Also note, in the same blog article:

Currently, fewer than 1% of search queries are affected by the site speed signal in our implementation and the signal for site speed only applies for visitors searching in English on Google.com at this point

[vandread]

I do not think it is such a bad idea but they have to be more transparent about it like notifying webmasters in Webmaster Tools if the site speed is affecting their rankings.

[TheMadScientist]

Yeah, IMO 'much ado about nuthin' ... Like I said previously, it's considered by many to be a relatively weak indicator, and IMO they know this, so IMO it's much more likely to be used against a site where the calculated load speed is way out of line with the calculated load speed of other sites in the niche, and people are getting all worked up over what right now is a non-factor, and will most likely remain a factor that's easier to use to indicate 'lack of quality' in an obvious situation, than a 'signal of great quality' to move a site up in the results. But, that's simply my opinion, so if others think it's a huge factor and want to concern themselves greatly with it, the feel free... My sites are all relatively fast, and my traffic has not been impacted, and the microseconds I could possibly gain by doing something different aren't going to change the calculation or relative speed, so it is what it is to me. I've got better things to do than try to explain how it's not IMO the worst of the factors (they used keywords for years didn't they?) and IMO speed will probably only have minute impact, unless your site is exceptionally slow.

[edited by: TheMadScientist at 10:51 pm (utc) on Apr 9, 2010]

[1script]

Currently, fewer than 1% of search queries are affected by the site speed signal in our implementation and the signal for site speed only applies for visitors searching in English on Google.com at this point

If it's so insignificant, why the announcement? Also, I would say that, given the enormous amount of long tail keywords out there, "less than 1%" can amount to many thousands of keywords.

[ken_b]

I just noticed in WMT that the cut off between fast and slow is 1.5 seconds. If I understand this right they are saying that 20% of pages (or is it a site avg number?) load faster that'n that.

I have a ton of nearly all text pages that probably beat that number, and another ton of pages with 1 - 5 images on each that probably won't come close.

Is my site fast or slow?

They say it's faster than 78%. How much will I be penalized, if at all, for that missing 2%?

I just want to build pages about stuff I like in a way that works for me and my visitors. Is that missing 2% all that critical?

[TheMadScientist]

If you're in a niche I compete in, pleeeeeaaaaase, spend every waking minute milking every last milisecond of speed possible out of each and every one of your pages, and pretty please, do not create any new content or get any new links or do anything else until you are finished. THANKS!

If you're not in a niche I compete in, I think you should make sure your doing the 'speed essentials' and are relatively as fast as the sites you're competing with, then move on...

If it's so insignificant, why the announcement?

To distract people, especially from all the bad pub. they've been getting lately? To me, it seems much less harmful to their business for people to talk about a ranking factor than all the copyright infringement and privacy issues in the news lately.

[aristotle]

Mad Scientist - My sites are relatively fast too, and most likely will be helped rather than harmed. But if Google is going to use speed as a ranking factor, even a minor one, they need to make sure that it's applied properly.

[TheMadScientist]

But if Google is going to use speed as a ranking factor, even a minor one, they need to make sure that it's applied properly.

I absolutely agree with you there, and I also think 'properly = consistently' so whatever the factors they use and whatever the mechanisms are they need to make sure they are applied so a site you have and a site I have that each load in relatively the same time generate the same perceived load time by their system. As long as they do that, then it should be fair and even.

Personally, I've used some of the speed checkers and they're not as accurate as I would like, because there were things I did it missed, and there are 'overall load speed tricks' I use sometimes to get things to the end user faster it told me were wrong to do, which is garbage...

BUT, I did a comparison using YSlow between here and a site I built that's 'screaming fast' too. (You can miss the refresh on a high speed connection if you don't watch the status bar either here or on the site I'm referring to.) Anyway, the reason for the comparison is WebmasterWorld loads 'straight through' (or did the last I checked) and mine loads a bit different but the overall 'viewable page, done loading' time is about the same, and even though YSlow said there were things I missed different than what it said about WebmasterWorld the percentages worked out to about the same (within 1%), which tells me there's a 'consistency' in the determination, even if not 'absolute accuracy'.

YSlow is also a fairly 'blind check' much like the one I would see Google as doing, so IMO even though they miss with the 'absolute accuracy' there is a 'consistency' between the results generated, which makes me not worry about the 'score' displayed too much... (At least there it seemed there was a 'consistency' in the sample of sites I checked. There were a couple that were off, but for the most part the score generated was about relative to the time to display.)

I didn't use the Google version of the speed checker, but if I was concerned I would probably start running it on competing sites, 'fast sites', and 'slow sites' to see if the 'time to display' and the checker's 'score' for each are consistent or not... Time to load goes up, score goes down. Time to load goes down, score goes up.

[edited by: TheMadScientist at 12:05 am (utc) on Apr 10, 2010]

[tedster]

With site speed currently a factor in only 1% of searches - done in English and only on google.com - it is not the likely explanation for the recent ranking drops that this thread was opened to discuss.

So let's discuss our ideas about Site Speed in this thread [webmasterworld.com] - and return this thread to it's original focus.

[tedster]

Well, a closer read of the new blog article puts some egg on my face:

We launched this change a few weeks back after rigorous testing. If you haven't seen much change to your site rankings, then this site speed change possibly did not impact your site.

So, if you DID see a recent rankings drop (is mid-January "recent"?) then you might want to check out your Site Performance stats. However, we should still keep the general comments about the new algo factor over in the dedicated Site Speed thread [webmasterworld.com] and use this present discussion for just those sites who saw mid-January ranking drops.

I also suggest reading the new article from Matt Cutts [mattcutts.com]. From his point of view "this is actually a relatively small-impact change, so you don't need to panic."

My gut tells me that most of what members here saw in mid-January is not related to the new site speed algo factor, and this present thread is much better spent comparing notes about other areas.

[myeuropeantouch]

1. Are your traffic Decline is from usa and UK?
USA
2. Is Traffic Drops only from Google or from other search engines too?
JUST GOOGLE
3. Did you notice any decline in GoogleBot Deep Crawling? YES
4. Is the traffic Drop is sudden or gradual decline?
DROPPED SUDDENLY MARCH OF THIS YEAR.
Is your rankings on your major keywords drops? YEA, TO
PAGE 8
6. Are your website internal pages caching normally?
YES
7. Decline in Indexing Status Of Your website?
YES

This is terrible, none of my store items show up in google anymore. What the heck happened?

[Webulletin]

First something about my site of around 40 products. Almost all pages were cached previously with better ranking. But situation changes since last 2 weeks. Here is my observation

1. Are your traffic Decline is from usa and UK?
- Yes. Global drop in traffic

2. Is Traffic Drops only from Google or from other search engines too?
- Only Google, all other are doing fine

3. Did you notice any decline in GoogleBot Deep Crawling?
- Yes.

4. Is the traffic Drop is sudden or gradual decline?
- Sudden decline

5. Is your rankings on your major keywords drops?
- Most of them (say 70%)

6. Are your website internal pages caching normally?
- No. I lost some of old cached pages also

7. Decline in Indexing Status Of Your website?
- Yes

[edited by: tedster at 8:54 pm (utc) on Apr 14, 2010]
[edit reason] removed identifying site details [/edit]

[maximillianos]

Oddly we have seen our site index ratio (number of pages we have versus what is in the index) drop from 85% to 35% in the last few weeks. We have around 70,000 pages.

Traffic has dropped by about 8%. So it appears they left most of our popular pages in the index, and removed many less popular?

Was that a goal of their new technology?

[drall]

Well nothing has changed for us since the mid Jan change.

Over a decade old site with zillions of backlinks dropped in indexed pages and lost 95% of it's googlebot activity with no changes at all to any part of the site,seo or server.

Speed? We are on the fastest backbone in the country and our server is sitting in the same room as some of the biggest sites in the world.

Site loads faster then 85% of the web so I dont think speed is the cause. Whatever has happened we lost quite a bit of our longtail traffic and it and the deepcrawl seem lost forever.

Bing on the other hand has every single page indexed and ranking within 24 hours.

PR7 site
Aged organic backlinks in such quantity it would make you cry
Handwritten high quality content
Super fast server, super optimized site
Ultrafast redundent global DNS
State of the art server security team

I give up trying to figure out wth is going on with them. Anexcercise in insanity.

[1script]

@drall:

I feel your pain! I have also lost longtail and still searching for what might have caused it. Nothing conclusive as yet. However, WMT account today finally updated the sitemap indexed pages count for most of my affected sites - and it's sharply down.

All the while the actual Google traffic is very steady at the new low level since the drop in March. In addition, site: count is steadily climbing up. Not fast but steady.

So, I have a theory and wonder if it can be supported by what you see for your site as well: the previous deep crawl has been "overly" reliant on sitemap data in terms of URL gathering. At some point in the past (Jan or March - both appear Caffeine-infrastructure related) something happened to the URL data from sitemaps that they kept for the affected sites. Say, a chunk of data got lost. The URLs they gathered via regular crawls survived. It may be a while till they re-crawl all URLs again - there could be any number of obstacles created by internal linking structure.

I can also speculate (and, again, this is just a speculation) that they might have "invalidated" URLs gathered from sitemap that they for some reason have not yet seen linked to via crawl. One of the possible issues I'm thinking of is the churning pagination dilemma - check this thread [webmasterworld.com].

Anyways, today is the first time in a month that I see the traffic drop supported by the data about indexed pages. Most (not all - I would be ecstatic if I can have the certainty of all ) affected sites show sharp drop in sitemap indexed URL count. Please check yours, post if you see something similar. Thanks!

[kidder]

Could it be as simple as a good old shuffle of the organic results = webmasters running to paid search = more revenue for Google. Now there is a nice simple equation.

[1script]

roll call:

How are your affected sites doing?

I am hearing from the Google April updates thread that there might have been another wave of G* traffic drops on or around April 16th. Any of your sites affected? Any recovered?

[cien]

I don't know what to think of Google anymore. Now they killed 70k pages from site:example.com in the last 3 days and enabled again the settings to set the spider crawling speed on WMT. They disabled it for about a week.

I had 4k pages indexed under site:www.mysite.com and now I have over 90K overnight even though my site is set to redirect www.mysite.com to mysite.com with vbseo for ages. I got now 20K indexed pages under site:site.com from 120K at the beginning of March.

Done trying to figure Google out.

[edited by: tedster at 1:20 am (utc) on Apr 20, 2010]
[edit reason] switch to example.com - it cannot be owned [/edit]

[Reno]

Done trying to figure Google out

Isn't that their point?

...................................

[Lorel]

1. Are your traffic Decline is from usa and UK?

No traffic decline in fact higher than usual. US.

2. Is Traffic Drops only from Google or from other search engines too?

I get about 98% google so don't watch the others.

3. Did you notice any decline in GoogleBot Deep Crawling?

I don't watch it.

4. Is the traffic Drop is sudden or gradual decline?

n/a

5. Is your rankings on your major keywords drops?

About the same.

6. Are your website internal pages caching normally?

Yes,

7. Decline in Indexing Status Of Your website?

Not my site but its' about 10 years old. Brand new sites take several weeks for Google to index them however, and to pick up new pages on sites not as old as mine.