Zombie Traffic from Google - Analysis part 2

[frankleeceo]

< continued from [webmasterworld.com...] >

based on your observations, does the zombie traffic hit the same pages over and over? or do they rotate within the site?

Google can and will toss different kind of foreign traffic against the sites for testing purpose. And I think based on user engagement and their social signals (or other signals), sites will have a higher affiliation toward certain countries, regions, or even their language settings over time.

My informational niche site has had regional traffic shifts in the past months, where I would lose traffic in one country but gain in another. Regional traffic may be at play too. Although it could be the searching trend of the underlying countries. But in a nutshell, all the traffic I lost in asia I gained the exact amount back in americas traffic.

[edited by: tedster at 3:47 pm (utc) on Dec 17, 2012]

[xcoder]

Last 6 Minutes... its been like that almost the whole day. Groups of smartphone users. Almost 85% of today's traffic were smartphones.

1) 202.124.75.4 [ No Reverse DNS record ]----[Netscape]----[Screen Size: 320x480]----[Color Depth: 32 colors]
Mozilla/5.0 (iPhone; CPU iPhone OS 5_1_1 like Mac OS X) AppleWebKit/534.46 (KHTML, like Gecko) Version/5.1 Mobile/9B206 Safari/7534.48.3
Referrer: Unknown [20:59:42] Reloads:1

2) 178.60.196.129 [129.196.60.178.unassigned.mundo-r.com]----[Microsoft Internet Explorer]----[Screen Size: 1024x768]----[Color Depth: 32 colors]
Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; Trident/4.0; .NET CLR 1.1.4322; .NET CLR 2.0.50727)
Referrer: [google.pt...] [20:58:54] Reloads:2

3) 86.1.241.146 [cpc6-rawt2-2-0-cust145.10-2.cable.virginmedia.com]----[Netscape]----[Screen Size: 480x800]----[Color Depth: 32 colors]
Mozilla/5.0 (Linux; U; Android 4.0.4; en-gb; GT-I9100 Build/IMM76D) AppleWebKit/534.30 (KHTML, like Gecko) Version/4.0 Mobile Safari/534.30
Referrer: [google.co.uk...] [20:57:27] Reloads:1

4) 202.124.72.31 [ No Reverse DNS record ]----[Netscape]----[Screen Size: 320x480]----[Color Depth: 32 colors]
Mozilla/5.0 (iPhone; CPU iPhone OS 5_1_1 like Mac OS X) AppleWebKit/534.46 (KHTML, like Gecko) Version/5.1 Mobile/9B206 Safari/7534.48.3
Referrer: [google.com.au...] [20:57:18] Reloads:1

5) 76.90.81.59 [cpe-76-90-81-59.socal.res.rr.com]----[Netscape]----[Screen Size: 320x480]----[Color Depth: 32 colors]
Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_1 like Mac OS X; en-us) AppleWebKit/532.9 (KHTML, like Gecko) Version/4.0.5 Mobile/8B117 Safari/6531.22.7
Referrer: [google.com...] [20:59:36] Reloads:1

6) 115.30.43.206 [115-30-43-206.users.engin.com.au]----[Netscape]----[Screen Size: 1024x768]----[Color Depth: 24 colors]
Mozilla/5.0 (Windows NT 5.1) AppleWebKit/534.52.7 (KHTML, like Gecko) Version/5.1.2 Safari/534.52.7
Referrer: [google.com.au...] [21:01:08] Reloads:1

[xcoder]

P.S.

I had an incline that Android phones are on par with iPhones or maybe even more. Apparently not, i see mostly iPhones for some reason...

I wounder what gives...iPhone tailored SERPs ? ?...

[claaarky]

A few posters are considering the possibility THE WEBSITE is the cause of this and I really think that's the right way to think about this problem (and many others!).

The conclusion of my own analysis (posted a page or two back) is similar to xcoder and gadget26, that this is an issue of our own creation. Google is mismatching traffic and from my own observations having watched traffic in real time during 'zombie' periods, searched on the same terms, looked at which area of the site people are going to most during those periods, looked at what types of devices were being used at those times and thought about why they are not buying, I've been able to narrow this problem to one thing.........the website.

Well, two things if you include Google's limitations - it has a much bigger problem understanding pages and how people interact with them than I ever imagined. It takes words from all over the page, the url, internal/external link anchor text, etc. and can send visitors on any combination of these words. Even words that WERE on the page once but aren't any longer (and haven't been for over 6 months), or words that are SIMILAR to words that WERE on the page 6 months ago.

It's made worse when that 'wrong' traffic actually spends a bit of time on your page (i.e. it's relevant enough to make people scan the page a bit but not relevant enough to make them engage with the site and buy). The result of this is Google thinks it's an okay match, so continues to rank you for that term. Add a great click through rate from Google, a surge of people searching on those 'wrong' terms, a naturally higher proportion of smartphone traffic at the same time and a dash of Google testing your pages with new terms (which may be more 'wrong' traffic) and you have the perfect zombie storm. Any combination of those factors will cause a noticeable zombie effect.

I really think the problem is pages that aren't tightly focused enough to help Google understand them more accurately, combined with the website's varying ability to efficiently deal with the traffic it generates depending on the device, browser, etc. that visitors are using.

I think there are parallels here with Panda issues which is why I suspect it's mostly Panda demoted sites that notice this (is there anyone here seeing this whose site is not affected by Panda?).

[Martin Ice Web]

claaarky,

but then big g has forgotten how to match the sites topic. Because prior to panda/penguin there has been no zombies.

I think its an algo /site problem. The question is, how can we convince the alog to recognize the right topic of a page?

[claaarky]

Martin, I've seen the zombie effect for years, well before Panda (since around 2008 I would say).

Traffic would rise, I'd get excited, but it wouldn't convert and would eventually disappear, sometimes along with a bit more traffic but then conversion would be excellent.

I think it was a problem before Panda and the sites demonstrating these problems (like mine) were then demoted when Panda arrived.

[Martin Ice Web]

Claarky,

I think it was a problem before Panda and the sites demonstrating these problems (like mine) were then demoted when Panda arrived.

That makes sense.
I never digged to deep into traffic before panda, cause the numbers were OK for me. But since panda the good traffic has been taken away and the bad zombies is all the has left. U would think it must be the other way, don´t you?

[TheMadScientist]

Martin, I've seen the zombie effect for years, well before Panda (since around 2008 I would say).

Hmmm ... That's when Google started changing infrastructures with BigDaddy, isn't it? Interesting.

Does anyone have info from back far enough to figure out when this actually started for them individually?

I actually think we're seeing the effect of a combination of causes here, so it's going to be tough to 'pin point' anything as what does it (EG it's 'not just mobile' and 'not just international' and 'not just previews or bots' from what I'm reading), but giving those effected something things to look for should definitely be helpful.

[edited by: TheMadScientist at 2:02 pm (utc) on Dec 17, 2012]

[Martin Ice Web]

The MadScientist,

funny, on last weekend i started to look at my old positions before panda/penguin went live.
So i search with the queries i hav from that time.
What i figured out is:
-low traffic sites with almost non conversions ranking better !
-my high traffic sites with lots of conversions got demoted or are not to be found in serps anymore

[TheMadScientist]

Okay, well if people are following this thread and they aren't doing a 'head scratch' over this one yet they should be ... We've got enough people effected and too much thought going into this for it to be 'nothing', but at the same time we've only got a limited number of members here actually reporting issues with it, and the dates of it starting on sites and plausible causes are all over the place ... Hmmmm.

BTW Thanks Martin

[claaarky]

Of course, the use of smartphones has increased significantly over the last year or two and it seems to be the 'norm' that even mobile versions of ecommerce sites don't convert half as well as the main site, so I think that magnifies the zombie effect for sites that attract poorly targeted traffic.

[Martin Ice Web]

Of course, the use of smartphones has increased significantly over the last year

if you use g keyword tool, u can look at data for desktop or mobile phones. desktop search is way in front of mobile.
Unless g uses very old data?

[backdraft7]

Not to break up the great conversation, but here's another Zombie Report from the midwest...
ON period, all day Sunday, until 5pm CST, then OFF, non interacting traffic ever since. It's now 10am on Monday morning, nothing yet. 17 hours of Zombies.

Can anyone correlate on that?

[tedster]

I'd like to re-focus us on the original issue. Our topic ihere s zombie traffic (not converting, not interacting traffic) that comes from GOOGLE SEARCH, and not zombie traffic in general.

In addition, this traffic seems to come in bunches. Some webmasters (not all) say that total traffic remains the same when the zombies kick in (we were calling this throttling.)

So generalized zombie traffic is really not the issue and that is clouding the discussion. Please, let's confine our analysis here to just traffic that appears to come from Google Search - thanks.

[Tyme]

Here's a timeline as best as I can remember.

As far back as 2007 our traffic was throttled. We are an adult/story site with some affiliate links. Our traffic was the same daily and then once a month it would go up by about 60 people and remain that way for a month and then go up again by 60 at the first of the month. It went on like this for years. We figured it was googles way of testing our urls. It continued this way until the MayDay update. Prior to Mayday we were getting a little over 2k people per month

The day after MayDay our traffic graph looked completely different, and our traffic was going up daily. By the time the first Panda hit we were getting about 10k a day and our traffic was converting on the affiliate links rather well. Also I feel we broke out of the traffic throttling because the traffic graph as displayed in WMT looked very natural, instead of the flat line, and then the spike at the first of every month as we had prior to MayDay.

The day after the first panda hit our traffic dropped by 60 percent and since then it has not converted well at all. (Zombie Traffic?) It's like a poor site as seen by panda gets the Poor traffic.

It has taken us over a year to make a partial recovery since the first panda. Interestingly after the partial recovery, we have also been making a little more money with the affiliate links.

personally I tend to agree with Claaarky on this. There does seem to be a mismatching of traffic for various reasons. Our Panda recovery has been partly due to the fact that we have gotten words off the page that could possibly confuse the google bot as to what the page is about. That and some other changes that have not been competed yet, I believe will get rid of the zombie type traffic and pull us out of Panda Land.

[backdraft7]

@Tedster - with all due respect, 90% of my traffic comes from Google. Zombie traffic IS non converting traffic so I fail to see how monitoring the real time situation is "clouding" the discussion. Observation of real world events and data go hand in hand and throwing out observation is severely limiting the overall analysis.
I'd rather tell if it's raining by looking outside rather than having to check the radar and rain gauge.

If someone spots a pattern and others can correlate, then we can come to some reasonable conclusions.
Both our data and observations are clouded by location, so gathering observations from many locations is key in unlocking this mystery.

I've said it before, I'll say it again...lotsa good theories here, but still nothing that really holds water. I'm currently deploying a CDN to speed up my sites and reduce the possibility that zombies are just impatient visitors who can't wait more than my old traditional 2 to 5 second page load time (or that G is penalizing my site on load time). Pages are now super fast, but zombie situation remains. Also now seeing a very odd pattern of traffic from specific parts of the US. Ohio valley has been very busy all day, with the rest of the country remaining deadski.
It's sad that my site is now more of a lab rat than a place of business.

The more I look at this problem, the more I suspect network issues, not just Google, but since both are a black box, we'll never know for sure. Theories are great, but without being able to prove them with testing, they will remain only theories. The trouble is, once you test & prove them, the game will change again... It's G's M.O.

[claaarky]

Backdraft7, I feel your pain and totally identify with the lab rat analogy.

However, I don't think a page speed of 2-5 seconds or network issues are the cause of your zombies.

I would look at the search terms for traffic that generates 100% bounces during your off periods, look at the pages those people landed on and the devices they used.

I think if you look at your pages through the eyes of the visitors who arrived on those search terms at those times with those devices, maybe even type those search terms into Google, look at your titles and snippets, click on some of the other sites ranking for the same terms, you'll see what's going wrong and why people are coming into the site but not buying.

Google wants your site to handle the traffic it sends with ruthless efficiency. That's not a game that's going to change. Things do change constantly but this is one that makes sense for your business and for Google.

Look at your top competitor's site as well (the one that ranks top all the time). They are not demoted by Panda and won't be seeing zombie issues to the same extent because their site handles traffic much more efficiently. They have teams of people crunching numbers trying to squeeze every penny out of every visitor. They'll have a great mobile site (or at least a much better one than yours). There's a lot to be learned from these sites.

[outland88]

I would agree with Claaarky that it goes that far back and he is likely right with the date but I would place it more mid 2009 before I started seeing what Backdraft and others mention. It’s frustrating and I never found the answer. I praise those who are sticking to their guns but I tired of the mods and their labels of conspiracy. Sometimes I think I have the answer but quickly return to doing anything that keeps Google in my back window.

[xcoder]

One of the more popular sites we mange for a client (pet related info site + active forum ) is simply drowning in a sea of smartphones... (Google referred "zombies"). Most Google ads on the pages are lately smartphone style/targeted.

So what do you think must have came first ?

Considerably more Mobile ads (them large ugly text blocks with large arrows that pay very little per click) OR considerably more smartphones viewers intentionally directed to pages that "accidentally" happen to display mobile ads.

P.S.
simple HTML pages + phpbb forum

[xcoder]

Saw the little critter in action again a few minutes ago. No "X_FORWARDED_FOR" or "Client IP" for this thing. Its an MSN bot masquerading as a browser. Look for it, it is definitely one of them frequent zombies. No "Bot" in user agent and can go easily undetected unless you do a reverse DNS check.

--------------------------------------------------------


131.253.24.40 [msnbot-131-253-24-40.search.msn.com] --[Microsoft Internet Explorer]--[Screen Size: 800x600]--[Color Depth: 16 colors]
Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.2; SV1; .NET CLR 1.1.4322) [10:33:22]

[Client IP: n/a] [X_FORWARDED_FOR: n/a]

Referrer: Unknown

[tedster]

Its an MSN bot masquerading as a browser.

How does that help us understand zombie traffic from Google Search?

conspiracy

Let's address this because it might help us understand what analysis is - and analysis is the topic of this thread. To say that "Google is trying to sell more Adwords" or "get the purchasing traffic for itself" is not an analysis. It is a comment that people make for almost anything Google does (Panda, Penguin, Knowledge Graph, etc, etc.)

But to be an analysis, any theory needs support from data. Along with others here and contacts elsewhere, I've been looking for that data support for a long time. So far, every time we think we found it, it evaporates on closer inspection. I'd love to break the news if we get the goods - but after all this time, many are very tired and bored with the same-old same-old that short circuits our discussions.

When every discussion we start ends up with "Google does it to make more money," then our discussions go nowhere useful. We get nothing that helps us with SEO or with establishing the truth of any theory. So we might not as well have any discussion if that's all we can say.

In the case of zombie traffic from Google Search, we don't even know for sure what we're looking at yet, to say nothing about the "why" of this apparent phenomenon.

Some members dug deep into their data and they saw that their zombie traffic seems to be mobile users. But we still don't understand why it turns on and off, for example. And others don't see zombie traffic showing mobile characteristics at all.

So we need to get some very disciplined analysis going to understand both WHAT we're seeing (KEY!) and then some idea of WHY it's happening. In that order.

[xcoder]

You are right tedster. Feel free to delete all my posts in this thread. I feel absolutely stupid sharing my data, knowledge and observations.

It is obviously not he right website for pros to share their knowledge... I'm out of here.

[backdraft7]

I guess I'm out too - to be brutally honest, I have yet to see the WebmasterWorld community uncover ANY of the myriad issues that Google has thrown our way. Let me tell you why...BECAUSE GOOGLE DOESN'T WANT YOU TO KNOW!
The second we figure out their modus operandi, they'll just have to change it again. I've said this a million times, but people prefer to keep looking. You have a better chance of discovering alien life or the meaning of the grand unified theory than to figure out how Google ticks.
WebmasterWorld is great for one thing, pouring our your frustrations and commiserating with other webmasters and unfortunately that's it. At this point I'm pretty much out of the game any, so have fun, happy holidays and all that good stuff. Feel free to sanitize my posts too. Good luck - C'ya!

[TypicalSurfer]

Tedster, the notion of google doing things for profit is not so far-fetched. I don't understand your insistence on defending the line for them. You're kinda losing some cred here by being absolutist about a situation that requires fluid thought. You've said that you don't even see the zombie phenomena across hundreds of sites that you monitor so what's the beef with others that do?

BTW, the zombie phenomena could be wholly related to adwords and how efficient players and their own algos have evolved to understand and react to live traffic. It's possible that "zombie time" is just that, periods of no or light commerce independent of paid or organic traffic. In any given market advertisers could be pausing campaigns during learned "zombie times" driving more traffic to organics.

Now that would be a neat explanation, it gets google off the hook for trying to make a profit and is plausible. I have no problem considering the possibility that advertisers are just as "self learning" as any other entity when it comes to prime time (ads on) vs. zombie time (ads off) and are simply reacting to something that has always existed. With ROI shrinking via adwords I know some people only want visibility during prime time, I'll pay for traffic from 9-11 AM every other Tuesday if that is what works and making things work is what the game is all about but pretending that adwords is not a factor in the game, well, that just isn't reasonable at all.

[Shepherd]

Tedster, the notion of google doing things for profit is not so far-fetched.

I'm sure I don't need to speak for Tedster but I don't think anyone has discounted money being a reason for zombie traffic. I think the point is: "the why does not matter until you figure out the how." Figure out the "how" and then go after the "why".

[TypicalSurfer]

Accusing people of entertaining "conspiracy theories" is just as distracting if not more so to solid analysis, it has built in connotations that only serve to engender more nonsense.

[SevenCubed]

I wrote this post last night out of shear frustration with tedster for once again jumping in attempting to suppress negative google sentiment. But at the last moment before clicking submit I was second guessing myself and decided to sleep on it. Well, I see this morning others are feeling the same way so here's what I was going to post...

To say that "Google is trying to sell more Adwords" or "get the purchasing traffic for itself" is not an analysis.

It's not an analysis but without bringing it (the motive) to the table then there is nothing to analyze. It's simple cause and effect. They are a for-profit company. Any changes in algorithm is with the motive to increase profits. Don't believe them when they say they aren't watching ad revenues, that's absolute nonsense.

When a crime is committed investigators typically cannot complete their investigation (analysis) without looking for a motive. How is that so much different from us looking for the money trail to determine motive for algo changes?

As an example...when I see local results of many highly qualified local businesses being pushed to page 2 or further (not just my sites) and the first page being littered with a cesspool of domain crowding by yellowpages type directories I have to ask myself why is that happening? Why are they willing to risk losing credibility as a good source of info? Then I can start my risk reward analysis based on knowing that whatever change is taking place has a financial motive. They will risk short spans of alienating their users for the reward of businesses buying more adwords. It's plausible enough for me so I put it forward in my posts.

What I'm thinking is they are taking a gamble that they can afford because of their monopoly. They appear to be squeezing small businesses for a while at a time to find out if they can force them into adwords. Then after a while they open the traffic valve again. Rinse and repeat.

So absolutely YES, realizing that they are trying to force more into adwords is a VERY credible observation. If you think it isn't then you are free to jump in and counter those observations by giving a counter-balanced opinion about why that can't be rather than just saying stop it?!

Yet I see far more members willing to offer strong arguments that support some of our observations than I see members willing to say it can't be that because...why?

So here's something to chew on as it relates to zombie traffic. They can surely differentiate between a mobile user versus a desktop user. I'm sure that they also know that most mobile users are most of the time performing search as a secondary action and that their reason for being mobile is because they have something more important that they are doing. They are not in buy mode.

Based on that it's not unthinkable that they can siphon desktop traffic more likely to convert and feed them the appropriate ads in SERPs at the right time and send non-converting mobile traffic to the businesses that they pushed to deeper pages by displaying them higher in SERPs for those types of queries from mobile users.

Another possibility is they can send potentially converting traffic from NYC to LA websites and vice-versa again for the purpose of making the ads more attractive.

There are many ways for them to throttle traffic for their benefit and for the benefit of the advertisers. I'm not saying they don't have the right to do so. I'm just saying it is not absurd to think they aren't doing something like that. Cloaking their nefarious intentions like that certainly can keep them below the FTC's radar can't it? This behaviour we are trying to dissect is unique to google so it is not related to crowd behaviour.

Only a few members may be experiencing this but maybe that's because they are unfortunate to be a test group. If it works out well it may become the new normal.

I barely care anymore and can hardly muster up the effort to discuss it. People protecting them and saying we cannot suggest financial motives certainly doesn't help the matter. That wears me out more then google themselves.

In defense of tedster I am aware that as an Administrator of WW you may sometimes have to moderate comments based on the wishes of the forum ownership. Tedster this is not a personal grudge, that's not the way I function. But I'm in agreement with others here about throwing in the towel. For you to throw the knowledge graph comment into this fray, it was not lost on me that was in reference to my flaming comments about it in another thread. That's why I am jumping in and responding here.

I'm outta here effective the end of today and will not attempt to participate again until January, if I still have an account to sign into. But I'll stop short of saying Adieu. I'm not being melodramatic, I'm just tired.

[Martin Ice Web]

I would like to add,that google just has an agreement with belgium press about using pieces of their news. The agreement says google pays a fee but the press has to take advantage of adwords to make their site more popular.

In a germany newsletter i read an article about the german negotiation to use the news from the news sits. google does not want to pay and defends against the home-standing law that search engines have to pay for the mental work of others. A german spokesman from newspaper union was quoted that google wants only the news to dosplay their ads but does not want to share their earnings.
I think in germany they will have no easy play with the press. We have laws for all and everything and have to pay a extra fee for CDs, copiers, DVDs, USB Sticks, Computers only because we have the ability to make a copy of something. google watch out.

I am sorry that this is not zombie related but is a thought about googles real intentions.

[netmeg]

Give up, tedster. I did.

[TheMadScientist]

Saw the little critter in action again a few minutes ago. No "X_FORWARDED_FOR" or "Client IP" for this thing. Its an MSN bot masquerading as a browser.

Huge Thanks! to xcoder for digging in and identifying this thing positively as a bot from Bing ... The 'it's Google doing it to make money' theory is developing a bit of a hole, since in at least one case, we have one of the 'active zombies' coming from Google's direct and determined competition, which I Highly Doubt is an effort by Bing to help drive advertisers to spending more money with Google ... I guess anything's possible, but that Bing's sending a bot appearing to be a browser (zombie visitor) in some 'covert, coordinated effort' between the two to help drive people to buying more AdWords from Google really isn't a conclusion I'm prepared to draw personally.

This discovery also makes me wonder:
What else are people missing?

[TheMadScientist]

I know this is a 'slightly OT request', but I'm wondering if tedster would consider allowing the identification of Anything Not from Google specifically in this thread, especially like the info xcoder found ... My reasoning is I think it might be valuable to allow people to 'eliminate some trees' from the 'forest' being looked at by 'shrinking' the Google/Unidentified zombie visitor source dataset through the process of elimination.

EG With xcoder's identification of an MSN Bot, those ranges of IP Addresses/User-Agent strings can be eliminated from analysis by anyone trying to figure out what's going on.

If he can identify what he saw and thought was a Yahoo Bot also, for sure, then those UA's/IP's can be eliminated from analysis too, and by doing so, it should help people to narrow down the dataset they're trying to find some 'Google zombie visitor pattern' in and if we do that, maybe we can get to conclusive/definitive reason behind it sooner.

I guess what I'm thinking is sometimes it's difficult to find a different conclusion or alternate reason for something when you look at it from the same angle all the time, so I'm hoping for a period of time in this thread we can 'change the view point a bit' to allow 'any and all definitive zombie visitor identification that's NOT Google related', because by doing so we should be indirectly 'zeroing in' and what's actually from Google and what's 'some other type of non-converting, non-interacting visitor' and that might actually help with analysis.