|Converting vs non-converting traffic from Google referrals and others|
| 9:33 am on Nov 18, 2012 (gmt 0)|
I’m following the Xcoder (SERP November changes) and Backdrafts (Zombie traffic) threat since creation date. Our established main sites are e-com (brand) and informational sites. Beside this, we run few more smaller websites.
We have 100% the same situation as Xcoder, Martin Ice Web and Backdraft. As we do have live tracking and own apps s for our server stats we can track real-time and in very deep detail. We noticed in the last 3 Month that useless (junk) traffic hitting our main websites, conversations and sales are down compared to the last year about 50%.
We droped by new created free accounts and sales …. but overall traffic is the same. Its just one day non-converting high - bounce vs. real...
Is some kind of G start to dance again (remember few years back) but this time its a robots dance vs real dancers it seems!
Our main site:
As special now in November our bounce went up from around 50% to 63%. (on the main e-com site). We did even change the layout and made the site much more user friendly, created a lot of pols what gave us a deeper insight in the user experience. Also our heat map shows strange (bot/robot like) behaviours by all the junk visitors (but the referrals are all from G).
We also noticed when we hit the G first search result page with new keywords, the traffic from other keywords is going down. In summery we always get the same amount of G referrals per month (steady). When a new keyword went on G first page we get a referral push from around 30% in total, but only for 2 or 3 days, then we went back to normal.
All this referrals have around 70% bounce. Its like a Zombie (not converting & very high bounce) attack. I know what you guys thinking, but its not that the site is not attractive in the SERPs, as every time after this kind of G mystery the new keyword gets established, stays on the same rank in the SERP and getting steady, but noticeable less traffic...
Its like at start we get the hole cake (what only zombies eating) and afterwards G splits the total (converting) traffic with other sites...
The November was even worse, it looks like we getting 75% zombies and conversations in first 2 November weeks were almost not existing (loosing big ).
The problem is that even in November ad-words is bringing non converting - high bounce traffic as well (ads are setuped correct btw...).
We got hit by penguin (main site) back in June and cleared our backlink-profile and did sent a disavowal file to G... We noticed that we recovered for around 50% of the keywords (that’s quite good news) but the over all traffic is the same (so G took instead traffic from other keywords) but this overall traffic is again junk.
We did noticed yesterday 18.11 a change in conversations. We got quite some sales in a couple of hours (was quite a buzzy day), like G is compensating our loss or stopped the dance....
No dance at smaller sites but still issues:
The smaller sites are around 1 year and 6 month old. Few sites are getting almost no visits form G (but we are sure there was no penguin or panda hitting) another site had backlinks issues and only got about 40 unique´ a day. Hugely has 3 keywords assigned. (site is 1 year old and was affected by panda, not because of thin content rather cuz of keyword stuffing in titles and duplicate title & description issues) but we made a total new site out of it 5 month ago... still traffic is not existing.....
Now in November the sealer sites finally starting to get some mickey mouse traffic. Over all the November was till yesterday, the worse month in this year!
We also noticed that G did assign and re-rank quite a good amount of keywords to our main site... few days ago … and the job G did seems to some degree OK for us, as we did target some of this keywords....but too much traffic dancing (zombies vs. real)
| 1:52 am on Nov 19, 2012 (gmt 0)|
It sounds like Google is sending you more mis-matched traffic now than before. For example, you could be getting relatively more traffic from keyword searches used by information seekers and/or "window shoppers".
| 3:09 am on Nov 19, 2012 (gmt 0)|
Not to give you a bigger headache than you already have but I just went from about a month of mismatched traffic to 50% less traffic overnight. My hunch is that mismatch leads to lower UX values which gets fed back into the ranking algo and that becomes a positive feedback loop.
Funny thing tho: about two days before the traffic dried up, they threw nice converting traffic at me. Lasted for about 36-40 hours, then half of it is gone, zombies as well as "normal" people that were coming in those nice 36 hours. And, BTW, what's still coming from Google can be described as "slightly warmed up zombies". Better than it was before, not even close to same season last year.
Don't have much analysis done yet- the trouble is very fresh, I'll need a few days to look at the data (and WMT will catch up with search terms data, too) to see which KWs did I lose. It feels like I've lost the "shorter end" of the long tail but it's hard to tell for sure yet.
| 3:48 am on Nov 19, 2012 (gmt 0)|
I would just like to echo TypicalSurfer in the "Google Updates and SERP Changes - November 2012" thread which i feel was spot on!
TypicalSurfer said: If you read the transcript from googles last earnings conference call they plainly stated that monetizing user intent was a big part of their direction.
Understanding their intent will help you understand the algorithm. Stop the magical thinking and the promotion of google as a search engine, they are interested in monetizing the worlds information, putting an ad overlay on anything they can.
You won't defeat the lion by pretending it is a rabbit.
Nothing to do with your website. They are clearly cherry picking our customers.
Their PR clown will hopefully confirm that soon... under oath.
[edited by: xcoder at 4:10 am (utc) on Nov 19, 2012]
| 4:01 am on Nov 19, 2012 (gmt 0)|
Without getting deep into my traffic, I do get a sense of traffic. Unscientific as it were.
I haven't seen a drastic decrease in traffic, but to me the site is pretty much dead.
To put it another way, the traffic isn't posting comments, they aren't click "+1's", they aren't buying anything. It's dead traffic and yes, I have that sensation. I see numbers in stats but the "feel" I have is that it's meaningless traffic.
In the summer, the same volume of traffic was doing things like commenting, clicking +1's, buying stuff, sending emails, etc. Big difference now and the numbers really aren't that far off.
Other than not spending more time on the site, I can't see a way to improve this. It's a very odd feel. Zombies, bots, or whatever you want to call it. It's complete junk and you don't need to look at stats to prove it. Look at any of the interactions on the site.
Another theory is that people coming to my site can't understand the language. It's the same way in which I'm getting so many foreign sites in my search result pages. Frankly this isn't worth much of my time. This situation isn't begging me to invest more time in a site that appears unable to convert right now.
Wake me up when it's over.
| 9:08 am on Nov 19, 2012 (gmt 0)|
I would like to add that looking to the live stats, we can see that the search terms entered in G are matching to 100% and our content is 100% related.
Can some1 explain why a G user what is searching for blue widgets and did found it, click on the result come to our site and close it in 1 sec? No interaction at all, but there are videos, pics, free products to download so on and all the info you can get about the blue widget + answers related to the search? Why would some1 do that? Maybe a small percentage, maybe.... but not 40 - 50%, that just doesn’t make sense! Put yourself in the situation: you search for blue widget (+ question to blue widget) you come to the 100% related site with all the info and you just close it?
And I do agree that since some months FB likes, G+ other shareing´s and even video views are down much! Hugely any interactions... if we wouldn’t still have sales and getting daily natural links (cuz of the high standard we provide, I would say the site is dead (but it actually isn’t as we have monthly 10 thousands of downloads and still sales as well...
But regarding to the unique per day we should have at least 30% more sales and other conversations... but with the traffic coming form G its just not possible... this behaviour we cannot see by yahoo or Bing... so it must be G related...
One of our coders think its some kind of attack to lower the bounce and user experience stats on our site, what you guys think? (but we talking form G referrals and we notice similarly behaviour with our ad words ads as well, but this is a another story I would like to come back later....
| 9:18 am on Nov 19, 2012 (gmt 0)|
^ prefetch fake hits (?)
The only way to verify is to run a "bot/human" test.
| 9:32 am on Nov 19, 2012 (gmt 0)|
Well it could be possible, but remember the entered search strings are not allays the same, from 100 visitors 80% have different search strings entered, it would need quite a research to know what keywords are related to our site (looking to GWMT there are quite few hundreds... and we getting traffic form them, who could copy this data and write a script what would target this keywords and run it OVER G (and G would block robots behaviour) with unique IP´s?
The searches look 100% real (well ofc some strings we do not see) but form the once we can see 80% are different and long tail (3 words min.)
This phenomenon is coming in ON/OFF rhythm, like Backdraft is monitoring in his zombie threat
If such a technique exists and could be done over G, then good night ...
Negative SEO in his purest form and most effective...
I personally don’t believe its an 3rd party attack...
| 9:37 am on Nov 19, 2012 (gmt 0)|
|Well it could be possible, but remember the entered search strings are not allays the same, from 100 visitors 80% have different search strings entered, it would need quite a research to know what keywords are related to our site |
Well not really. Googel hold quite extensive search history for each and every website, including keywords, going a few years back, ...
I see Google bot often entering search keywords in some of my websites search box and carry a search.... very common actually.
They cache and store everything via their numerous spyware tools, including user interaction over your very own website.
[edited by: xcoder at 9:56 am (utc) on Nov 19, 2012]
| 9:41 am on Nov 19, 2012 (gmt 0)|
"write a script what would target this keywords and run it OVER G (and G would block robots behaviour) with unique IP´s?" and unique hosts ...
That would need a mega effort and budget (if possible), no? I don’t think we are so important...
| 10:57 am on Nov 19, 2012 (gmt 0)|
The Google algorithm is messed up, causing a lot of mis-matched traffic to be sent to your site.
| 11:35 am on Nov 19, 2012 (gmt 0)|
how it can be miss-matched traffic if the search string is 100% related?
I´m missing something?
| 12:31 pm on Nov 19, 2012 (gmt 0)|
|how it can be miss-matched traffic if the search string is 100% related? |
Google now hides the keywords in search strings (unless it is paid ads traffic)... how can you see organic search keywords strings? ...please elaborate.
| 12:42 pm on Nov 19, 2012 (gmt 0)|
The string is hidden by user what are logged in in their G profile and as far as I know youtube strings are hidden... the rest is of course visible (don’t know about G+ tho)...
I see strings from all G search properties world wide (.com, .au. uk so on.. except from the G services what do hide the strings... but that’s by far not all...
Do you think G is hiding ALL strings form all his properties? I do not believe this and I have evidence, except there is something other going on what does FAKE G referrals.. but I doubt that...
| 12:46 pm on Nov 19, 2012 (gmt 0)|
i would say it´s 50/50 of hidden and visible strings... something like this
|Martin Ice Web|
| 2:51 pm on Nov 19, 2012 (gmt 0)|
nice post. Good summary of what we see since panda/penguin april.
The most important thing I see is, that traffic for each website thats been hit by panda has a limited amount of clicks. While this clicks have a limited amount of real traffic. This, like you say, is totaly independend from amount of pages you add to the page. Keys are rotating through the websites different targeted sections.
The most wondrously thing is, that pages on search page #1 will get no significant traffic. This is, due to my reasearch, a result of googles similar context search. Google just ignores a key or replaces it with a not matching keyword out of its "similar database". ( only big brands, amazon and ebay seems not to bee affected by this similar error ). I think this is the zombie, not converting traffic, that will produce a high bounce rate.
Common (glabal ) 2-3 word searchs will show your site, adding a deeper keyword, will not show your site. +
Like "my big yellow widget" shows up
"my big yellow widget 5m" will not show up
This, imo, is owing to the fact that panda/penguin is although a keyword overoptimization filter. The more your site matches the query the more negative "panda score" will be used to place the site in serps.
Also in this case brands, amazon and ebay are not effected.
Since last panda updated i see old black hat spammers coming up in serps again. Using new black hat technics. Stupidly, because of pandas own sight of "relevant content". This sites do have no navigation, only one, two links, little html, contain the key in h1, title, meta tags, but less in page content.
| 3:31 pm on Nov 19, 2012 (gmt 0)|
@ Martin Ice Web
Thx for your opinion I appreciate it!
"each website that’s been hit by panda has a limited amount of clicks"
agree to 100%! No matter what new content we burst out - over all traffic is the same... its like "GIVE MORE HERE and TAKE SOME THERE ...
"Google just ignores a key or replaces it with a not matching keyword out of its "similar database"
That would make senses: Seeing a lot of 2 word strings and prolly G was cutting out a 3 rd or 4 th keyword in front or after the string.. so its possible the searcher was searching for something other but it was related to 2 words of our SERP listing...
Still we have very descriptive descriptions, stating clear what’s going on on the specific post, site so on...
Based on your theory and observations, the searched would ignore the description click and come to us, but how he can verify in 1 sec that it isn’t the information he was searching for.. all our pages have up to 1500 words, videos, and pics ...
I think it must be something else what causes the high bounce, even i do agree with the theory you have to some degree...
We get with almost every new content on the 1. search page somewhere between position 1 - 7. it might be that if NO panda penalty is given we would always be on positions 1 but still we archiving good results and are very present in the SERP with descriptive titles and descriptions.. so this is the mystery I try to solve... why some1 would click and leave in 1 sec.... it takes some time to the searcher to summarize the page... i would say 2 - 3 sec at least (if its a spammy site or ugly or what ever, but this is not the case with our sites)
"less in page content."
There is a theory going on that panda watches on coding, maybe we are all bad coders or make too many mistakes and the spammer doesn’t have this issue as he provides no content :-)
| 3:44 am on Nov 20, 2012 (gmt 0)|
|The more your site matches the query the more negative "panda score" will be used to place the site in serps. |
Those Google engineers must be smokin' some bad 'chit these days if that's really what is going on here.
I agree about the Amazon/Ebay being somehow exempt... $$$?
Regarding blackhat tactics, I'm seeing TONS of facebook note posts spam showing up in the serps. And guess what, they're affiliate Amazon posts!
Lean pages... very little content pages seem to be doing well..... guess, they're passing the Google Hookah again.
We're in the slow process over converting our site to HTML5, but we are having to be very cautious because we still show users with much older browsers.
Interesting theory on the pre-fetch being the cause of some of the zombie like "traffic" Is it possible that Google actually counts a pre-fetch as a bounce?
Ok, someone pass me the pipe so I can understand.
| 3:54 am on Nov 20, 2012 (gmt 0)|
Bewenched, it's almost as if Google is filling 9 of 10 serps with the "Feel Lucky" function.
Last week started fine, but then slowed and totally shut down on the 18th and 19th.
What a roller coaster!
BTW - US Traffic is appears to have suddenly gone Zombie. Two sales today, one from the UK, one from Canada. This holiday week is typically very good sales wise.
| 4:58 am on Nov 20, 2012 (gmt 0)|
I guess the big brands are getting the real traffic and business. Shame that the smaller businesses don't get their fair share... you'd think Amazon should also be relegated to Google Shopping only like the rest of the ecommerce sites, goodness knows they have plenty of money to spend on ads and affiliates that spam the web. And they have duplicate products and content all over the place.. the whole site is full of it. Why not make them pay to play.... or maybe they are.
I remember a few years back you could boost your rank in Yahoo by paying to be on the first page of results, which absolutely appeared to be a natural result.
I wonder if behind the scenes Google isn't doing the same thing.... it's really starting to look that way. They are a publicly traded company and it would certainly be within their right to "tweak" results to suit their bottom line and share holders... nothing unethical about doing that... honesty, well that's something else.
Maybe we should all just turn into Amazon affiliates and sit back and make small percentages of sales, have no staff, no techs, no phones. Or better yet, just outsource all our customer service to India, host millions of sites that scrape other peoples content and images and don't pay our state sales taxes where we have warehouses. (yea, amazon does all that and more) Why would Google give such preference to a company like that.
I hope this latest "shake up" passes soon, our conversions were decent until about 4pm central time today, then nothing since. I'd love to know why we get this zombie traffic.... and why sometimes in google shopping it shows two clicks for a product yet only ONE page view.
Sorry for the rant, it's been a long hard couple of years and this year has had it's share of sleepless nights battling bad bots, updating source code and it just seems all for nothing.
| 11:24 am on Nov 20, 2012 (gmt 0)|
our conversions were decent until about 4pm central time today
I noticed a G datacenter swich