Ranks Remain Stable - Google Traffic Halved

[simonmc]

So whats this then. Over the weekend, something tanked badly. Last week was one of the best weeks ever ranking and orders wise. Then... BOOM! or should I say BUST!.

Across datacenters ( I have even checked using several proxy servers) my rankings remain where they have been but the traffic has halved. I'm talking about a site here that usually gets 2000 uniques per day getting 1000.

How is this even possible? How can google still rank as it has done but traffic from google halve?

This makes no sense to me at all.

[BigDave]

Did you compare your ranking on EVERY keywword that brings your traffic? A site of mine that gets similar traffic to your's has about 25k keyphrases each month that brings in traffic, I can't imagine trying to keep track of all of them.

[surfer67]

My Google traffic also halved this weekend. First time this has occured in the 6 years I've had the site. I'm still listing high in the main keywords. No way to check the thousands of keywords that bring me traffic.

Hope it's just a hiccup and stabilizes soon.

[simonmc]

I am tracking around 30 keywords. These all appear stable. I have checked these using many proxy servers from around the web so I am hitting google from loads of locations. All holding up well.

I also checked to see that I was still in MSN and Yahoo though to be honest they are only responsible for about 20% traffic.

Anyway, I too hope this is just a glitch. I know of one other person that has had a big traffic drop too that happened around the same time.

Anyone else?

The key is that rankings are stable yet traffic has had a significant drop since this weekend.

[Tinus]

Is the problem maybe that the long tail searches are halved and not the general searches with 'moneywords'? That would be the oposite of what usually happens first when you trigger a filter.

[edited by: Tinus at 3:39 pm (utc) on Oct. 24, 2006]

[johnhh]

This problem/odd effect was spotted and noted over at [webmasterworld.com...]

Nothing odd in the serve logs just a lack of volume

In the process of checking quite a few hundred known keyword combination SERP positions..

OK I confess I've got the wife to do it bless her.

[texasville]

>>>>>How is this even possible? How can google still rank as it has done but traffic from google halve? <<<<<

I had the same thing happen over night to my main site. It's smaller tho and I can track it better. In the last two months it had risen to the top five and traffic was building. Now, it's halved. The funny thing is it's almost exclusively three word search terms that have disappeared. It is just gone from there.
I also noticed that one term which I ranked well for and now gone was returning 17 mil in search and now is down to 12 mil.

[BigDave]

simonmc and others,

I think you just got a lesson in how powerful the long tail is. Even if those 30 are your top 30, they are still probably a small portion of your traffic.

In the case of a site that is averaging 4,000 visitors a day, my top 30 keywords account for 0.46% of my search engine traffic. I'm sure that your top 30 provide a larger percentage than that, but you best take a serious look at your stats and figure ouw what percentage they really represent.I would be shocked if you get those sorts of traffic numbers and your top 30 represent more than 20% of your search engine traffic.

[simonmc]

So what could cause the longtail to get hit and not the main keywords?

[jetteroheller]

The symptoms at me are the same. Much less Google traffic, all major search terms okay.

I also checked 20 referer log entries from October 14, all in the same position.

Because I compare each day my stats with the total stats of
[webhits.de...]

I have much more exact data about the developoing

Data Ratio Webhits.de to my page impressions

Montag..........2006-09-25...145,15
Dienstag........2006-09-26...122,77
Mittwoch........2006-09-27...118,59
Donnerstag......2006-09-28...126,72
Freitag.........2006-09-29...140,22
Samstag.........2006-09-30....82,61
Sonntag.........2006-10-01....80,81
Montag..........2006-10-02....86,78
Dienstag........2006-10-03....76,16
Mittwoch........2006-10-04....79,02
Donnerstag......2006-10-05....80,70
Freitag.........2006-10-06....83,59
Samstag.........2006-10-07....88,55
Sonntag.........2006-10-08....83,55
Montag..........2006-10-09....88,45
Dienstag........2006-10-10....86,74
Mittwoch........2006-10-11....91,61
Donnerstag......2006-10-12....90,95
Freitag.........2006-10-13....94,96
Samstag.........2006-10-14....95,91
Sonntag.........2006-10-15....86,99
Montag..........2006-10-16....90,71
Dienstag........2006-10-17....88,97
Mittwoch........2006-10-18....90,87
Donnerstag......2006-10-19...105,73
Freitag.........2006-10-20...104,84
Samstag.........2006-10-21...128,24
Sonntag.........2006-10-22...128,51
Montag..........2006-10-23...150,21

This shows the effect started strong at October 19th

I hope this strange effect becomes weaker, todays stats look half way to recover.

[jetteroheller]

So what could cause the longtail to get hit and not the main keywords?

I tested with 20 radomly taken referer log entries from Octover 14th

All my search positions are okay, the question is only on how many data centers.

[BigDave]

Well, I'm guessing you have concentrated your efforts on your major search terms, so they have lots of factors working in their favor.

On the other hand, long-tail terms are the ones that rank in spite of your lack of effor on them. Their ranking is not as robust as your major terms that you have concentrated on.

Let's suppose that it is a simple 2 factor algo, PR and anchor text. You have oprimized your major terms with lots of anchor text. There is little or no anchor text for your long tail, but they still have the PR and rank better than some that even have some anchor text but lower PR.

Then google turns the knob and increases the importance of anchor text and decreases the importance of PR.

You main keywords are sitting pretty, because they have anchor text. But your long-tail starts to suffer if the keyphrase has some competition with some anchor text.

[martinibuster]

This probably isn't applicable for everyone, but if US traffic is a factor, don't forget there's a Baseball World Series going on. My weekend traffic was significantly lower because of that and football.

Monday was within the range of normal and today is too.

Hobbies and business traffic was down. Travel sites had unusually higher traffic.

[edited by: martinibuster at 7:17 pm (utc) on Oct. 24, 2006]

[jetteroheller]

Well, I'm guessing you have concentrated your efforts on your major search terms, so they have lots of factors working in their favor.
On the other hand, long-tail terms are the ones that rank in spite of your lack of effor on them. Their ranking is not as robust as your major terms that you have concentrated on.

Resarch of a major search term, where I am between rank 3 and 6

Sonday 15.....36
Monday 16.....30
Tuesday 17....48
Wednesday 18..29
Thursday 19...40
Friday 20.....43
Saturday 21...28
Sonday 22.....15
Monday 23.....23

[simonmc]

Surely if Google has tightened up filters on the longtail searches, then this forum would be flooded with "My traffic has halved" kind of threads. Rather than this pretty pathetic stream of just a few people.

That does not make much sense does it?

[Tinus]

It is strange indeed. But you can check in your statistics maybe what kind of search words are lessened and confirm is long tail serches are lacking. Then we can hypothise about possible causes why some are hit and most didn't. My traffic went up this week so I can't check what you might experience.
If it is true that long tail searches are lessened for you, the post of BigDave makes sence as a possible reason.

I wouldn't use the word pathetic. All kind of information is valuable to identify a problem. When we know exactly what is the problem finding a solution might be easier.

[sailorjwd]

I've noticed a dip in traffic recently too.

I took at look at the Google sitemaps page where it shows popular keywords that folks use which show your site and ones they use where they actually visit the site..

In the past I had many #1 and #2 positions. Now I see no #1's and very few #2 positions.

In my case it may be a temporary screw up - I added about 10 words of text up at the top of each page. This may have diluted my keyword density.

[johnhh]

its certainly worth investigating as may be the first instance of a change that will affect more later.

[jetteroheller]

Let's investigate the sites with this strange stable rank, google traffic halved.

Only my main site and the themed subdomains of my main site is affected.

Very stable site since April 1997.

One problem middle December 2005 with a scrollable DIV so small, that Google told me "hidden text", problem resolved December 31th

June 27th 2006 suddenly most sub domains filtered, 85% traffic loose
July 27th some sub domains out of filter, a few more in filter.
August 17th all in filter
September 30th all recovered

[trinorthlighting]

Have you checked your competitors? May be they changed their description tag and it attracts more people. Might not have anything to do with rank, could be your competitors descrption in the serps.

[pageoneresults]

Maybe off topic but I just found out that Comcast customers that were forced to migrate over to the new Time Warner interface on the 20th of October don't have access to quite a few sites. I know, I just spent almost two hours with tech support trying to figure out why I couldn't access a client site.

It has now been 5 days without access to that site and Time Warner states they are working to resolve this issue. This affects a very large portion of Los Angeles and parts of Southern California. I'm sure this is affecting many. Some of the sites on the list that can't be accessed are ebay.com, target.com, etc.

[tedster]

There is also the possibility that DNS Cache Poisoning [webmasterworld.com] or related sabotage is siphoning off a percentage of traffic.

[SuddenlySara]

pageoneresults makes sense as well as a few posts up above.
about 3 years ago I had Time Warner internet in Los Angeles and was forced to update to Comcast (the other way around) and could not access many major sites for about 2 weeks or so. I'm south of LA now in Orange county and still with Comcast so no changes here (I can't believe areas in LA are now back to Time Warner?).

The World Series of baseball, basketball getting underway along with hockey and such keeps people away from the internet.

Also big changes in Goooogle searches maybe affected your long tail searches. I made it back on top this weekend on MAJOR keywords and long tail searches changed quit a bit.

[jomaxx]

I noticed a minor but noticeable drop this past weekend. All my rankings appeared solid, so I kept digging. Turns out pretty much everything was unchanged except that my google.co.uk traffic tanked.

[jetteroheller]

There is also the possibility that DNS Cache Poisoning or related sabotage is siphoning off a percentage of traffic.

This is a very good theory

I have now experiences with 2 Google traffic drops before.

The symptoms are always a great reduction of AdSense eCPM.

Visitors from Google have a much higher CTR than the other traffic from repeated visitors.

This time, the CTR and eCPM remains stable.

So it seems this theory could be something to investigate.

[johnhh]

results of keyword investigation

sampled 647 keyword combos we have historic data for 179 moved 10+ or more places down the SERPS, 40 moved up 10 or more places.

Traffic still down 20%, no real change noticed for competitors, no relationship to cache dates, and not related to pages design as pages with same design and same keyword density move up and down the SERPS.

Is this a issue related to sites that are not .com's I wonder, although possibility it is related to inbound links being discounted rather more than in the past.

[jetteroheller]

Just checked again with my referer log files about 20 queries from Friday October 13th.

All positions hold, but traffic halved.

No idea what this is.

I checked with Google Analytics

14..20 October
4240 visits 12166 page views
77.1% from Google 10.2% direct

21..28 October
2290 visits 7072 page views
73.8% from Goolge 12.6% direct

this compared to the time where this site was in a filter
with about the same total impressions

9..15 September
1314 visits 4190 page views
36,5% from Google 36,1% direct

And this is strange: when Google sends only half the visitors, the % of visitors from Google should also decrease dramatic like in September, but here only 2,3% less.

It's now like all the traffic is reduced.

[tedster]

Here's a suggestion for people seeing this footprint of "stable rank and traffic halved". Do one of the searches several times an hour for several hours. Look at two things:

1. Is the SERP position relatively constant?
2. Actually click on the result each time -- does the click go to your URL?

If you can get a few people to help with this experiment from different locations, that would also be good.

[Tinus]

Strange indeed.
The direct visitors should stay the same in absolute figures so it should relatively grow when you have less Google visitors. I assume your site is not about something seasonal?
When all of your positions hold, which kind of words are tanked then?
Are word combinations which are not directly anchor text or wordcombinations which are not part of a link to your pages tanked? Could you do some inventarisation?

[goubarev]

Not sure if it'll help or not: But I don't see the decrease in traffic in any of my sites... this is across few very large, large, medium and small sites in few different industries...

To simonmc, you know that traffic came from google, you know the possitions for main keywords didn't change so
- either positions for "other" keywords changed (the long-tail was lost)
- or either less people went to Google in general (not likely - unless you have "baseball-specific" crowd...)

That's my 2.5 cents :c)