Forum Moderators: Robert Charlton & goodroi
All our sites used to get heavy traffic from google..
We implemented google analytics on site 1 which used to get more then 1,00,000 uniques from google alone.. After 3 days traffic was dropped down to 5000 and 1 week later just to 1200 uniques a day (we sufferred a -50 penalty here)
on few other sites we included google custom search (paid) site traffic was dropped by 50% in a months time and still the same.. Our content was indexed in nearly 15 minutes on this site but now it takes more then 2 hours for same to get indexed.
Strange and confused :)
There is absolutely no correlations between the two. I have done this on many websites and seen no issues at all.
What may be true for your web sites may not be true for everyone else's sites. It is unlikely that statistically you have enough web sites on your own, with identical variations to the population of web sites in Adsense to have enough of a sample to know for sure.
It is unlikely that statistically you have enough web sites on your own, with identical variations to the population of web sites in Adsense to have enough of a sample to know for sure.
Nobody can be sure of anything, but the "common sense" rule applies here:
To echo ecmedia, why would Google want to punish Web sites for using its products? If Google wanted to be arbitrary, wouldn't it make more sense to punish sites that didn't use products like Google Analytics (and which might have something to hide)?
but the "common sense" rule applies here
Common sense is generally not universally agreed upon. To me common sense might be that Google would use the analytics data to calculate Adsense earnings. To someone else, it may mean something different.
Personally I would never claim to know what they actually do or do not use, since I don't have the data to support a conjecture either way.
Google would use the analytics data to calculate Adsense earnings
If you use adsense code on a page, Google already have access to all the visitor and referral information. They don't need GA to get that.
There is absolutely no correlations between the two
Not to be pedantic, but this requires clarification. Some webmasters certainly seem to have seen correlation (i.e. a GA install coinciding with lost traffic). This is to be expected with the huge numbers using GA, and the frequency of change in Google results, and even with adsense pricing. It must happen pretty frequently as site's search engine traffic is often unstable.
But correlation is not causation: I have seen no evidence whatsoever to even imply that there could be a link between the two - whether the two events coincide or not. I saw a drop in traffic the other day at the same moment it started raining heavily. Even if the same happened tomorrow I would not be looking to this as a cause ;)
It is unlikely that statistically you have enough web sites on your own, with identical variations to the population of web sites in Adsense to have enough of a sample to know for sure
Yes, but this cuts both ways - most reports of traffic drops and other side-effects related to GA are from individuals looking at a small numbers of sites, and usually without considering the other factors involved: choosing instead to opt for events that coincided. If there was a genuine link between GA and performance, we'd be seeing a lot of reports from a lot of reputable sources, since GA is used by millions of websites. In the absence of that, a causal relationship is highly unlikely at best.
Those claiming a link need to come up with evidence to support this theory - the onus is on them to do so. It doesn't require other people to 'disprove' it. The only evidence I've seen so far points completely towards correlation, not causation.
The only evidence I've seen so far points completely towards correlation, not causation.
And for some of us, nothing points toward correlation. Quite the opposite, in fact.
Why would Google hate you and love me?
I can think of any number of reasons why Google might hate some sites while loving others, but those reasons have more to do with the sites (or possibly with their SEO techniques) than with Google Analytics.
Those claiming a link need to come up with evidence to support this theory - the onus is on them to do so. It doesn't require other people to 'disprove' it.
The same logic hold true both ways. Absence of proof is not proof of absence.
[edited by: Jane_Doe at 4:16 pm (utc) on July 16, 2008]
The same logic hold true both ways. Absence of proof is not proof of absence
But absence of proof doesn't make a conclusion more likely, and is not an argument in favour of something being either true or untrue. Absence of proof does not make it impossible that something is true, but it often makes it less likely. The bolder the claim, and the less likely it is to be true, the more evidence (or 'proofs' if you will) required to support it.
It would be relatively easy to collect evidence that showed a causal relationship between adding code to a page, and instant decreases in search traffic. In the absence of any such evidence (and instead with only a small volume of anecdotal evidence of a correlation) the claim is significantly less plausible.
The person that makes such a claim has to provide the evidence - not the other way round.
I rarely use (and don't even particularly like) GA myself. I have no vested interest in one claim or another about it being true, but thus far no-one seems to want to provide any evidence to support their claims about it. I've seen a relatively large volume of sites that have added GA without any side-effects. Until someone comes up with something more reasonable I'm going to remain unconvinced ;)
[edited by: Receptional_Andy at 6:27 pm (utc) on July 16, 2008]
GA-CS >> Drop
The contrapostive of any true statement is always true, and the contrapositive of any false statement is always false. So if the above proposal is true, then its contrapositive version must also be true. The contrapositive statement says "if there is no rankings drop, then the site did not just add GA or CS." I'll use ! as the symbol for "not"
!Drop >> !GA-CS
Because a statement and its contrapositive are logically equivalent, just one example that contradicts the proposed idea or contradicts its contrapositive is all it takes to DISPROVE the propoal. There definitely are counter-examples to this idea, some of them are posted above. So a cause and effect connection is disproved, and some other factor is involved in the rankings drop.
The entire concern about GE and CS here is parallel to the long lived myth that adding Adwords hurt your organinc rankings - or that it helps your organic rankings. Both versions of the idea lived for much too long. I hope this idea dies out more quickly than that.
So a cause and effect connection is disproved, and some other factor is involved in the rankings drop.
The only thing this thread may suggest is that adding GA does not cause a drop in ranking 100% of the time for 100% of the people posting. For example, if in reality suppose for 10% of the web sites with certain factors adding GA caused a drop in rankings specifically due to adding GA, then it may be expected that most people would post no correlation but a small subset may post about a drop.
You can't prove there are no black swans (a drop in rankings due to GA) no matter how many white swans (no drop in rankings based on GA) are posted.
Personally, I've seen a enough of these threads to think twice regarding GA, though I would never claim to have proof of either position.
suppose for 10% of the web sites with certain factors adding GA caused a drop in rankings
Then the question becomes as to what those other factors are, and what significance they have. And of course, once sites have issues that can cause a significant ranking loss (whether 'triggered' by GA or not) cause and effect is even murkier: such sites are liable to suffer ranking drops at any point, and in conjunction with many unconnected activities.
You can't prove there are no black swans (a drop in rankings due to GA) no matter how many white swans (no drop in rankings based on GA) are posted.
Such an example does not apply to a causal relationship: it's not existence but causality that is the question here.
A better way of expressing it might be to propose that "if swans eat plants they become ill". You only need to find one example of a swan eating a plant and not falling ill for this premise to be incorrect.
We're likely to end up with a statement like "some swans with pre-existing conditions that eat plants might fall ill". And of course, by that point, the plant-eating has pretty much ceased to be the issue of any importance: what's wrong with the swans that suffer unusual side effects is, and it might turn out that eating habits had nothing to do with it - it was just the pre-existing condition.
If you do not add Google Analytics (GA) or Custom Search (CS) to your site, then obviously there can be no impact one way or the other on the algo.
If you DO add it and the data it feeds into the algo has a negative impact, then you'll presumably suffer; if the data has a positive reinforcement, then you may benefit. In either case it is not GA or CS specifically, but how the algorithm interprets the data they generate.
Personally, if I added either GA or CS and thereafter saw a steep decline, I'd pull it off ASAP, just in case that was the trigger for the drop. Maybe it was and maybe it wasn't, but I'd err on the side of caution until convinced otherwise.
................................
If you choose the anonymity option in GA then the collective data that may be used to tune the algo is anonymous. It's preposterous to think that the algo is tuned to target any particular website out of the millions/billions that are online.
Either it's anonymous and not correlated with any particular site, or else they are flat out LYING, including Matt Cutts, who has stated that GA data is not used by the webspam team.
I've been on WebmasterWorld and this forum for eight years, and in every single one of those years there has been some flavor of excuse to put the blame for lowered rankings on anything other than flaws with the site itself.
<yawn>
This is just the latest excuse; others have gone before and others will come in the future.
</yawn>
<aside>
Andy, hats off you to ya mate, for uncovering yet another ranking factor. Your comment about the rain correlates with my recent experience, so it must be true. Since the weather here in Calif. has hit the high 90's recently, I've seen rankings go UP for some sites.
That's it! They're using weather conditions to tune the algo and obviously it doesn't have a geographic component or we wouldn't both be seeing the same exact cause and effect correlation. Therefore, it's a fact!
</aside>
It's preposterous to think that the algo is tuned to target any particular website out of the millions/billions that are online.
Are you saying that you know with 100% certainty that information from Anayltics such as bounce rate, time on site, etc. are not passed as factors to the ranking algorithm?
Since the weather here in Calif. has hit the high 90's recently, I've seen rankings go UP for some sites.That's it! They're using weather conditions to tune the algo...
Maybe global warming will contribute to a rising tide that lifts all boats and Web sites?
Are you saying that you know with 100% certainty that information from Anayltics such as bounce rate, time on site, etc. are not passed as factors to the ranking algorithm?
They use the data in aggregate, which they plainly state; and there are levels of disclosure that are under the direct control of the user.
[google.com...]
[google.com...]
Q: Will sharing my data directly affect the ranking of my natural search results, ad quality score or ad placement? [Back to Top]A: Your website data will not be used to affect your natural search results, ad quality score or ad placement. Aggregate data across many customers will be used to improve our products and services.
My guess is that they ran that by their legal people in conjunction with disclosure and honoring their privacy policies.
[edited by: Marcia at 11:44 pm (utc) on July 16, 2008]
It's preposterous to think that the algo is tuned to target any particular website out of the millions/billions that are online.
If you choose the anonymity option in GA then the collective data that may be used to tune the algo is anonymous.
So.... sandyk20, did you choose anonymity?
.....................................
I can only say with 100% certainty what I read in their public documentation - which is either truthful or not truthful.
Thanks for posting that. I had not read that. I still probably wouldn't use it because I'm overly suspicious, but using Analytics based on that makes more sense than just assuming they wouldn't use the data based on one person's experience with their own sites or conclusions based on postings on forums.
[edited by: Jane_Doe at 11:58 pm (utc) on July 16, 2008]
and risk blatantly violating their publicly stated policies.
That's another compelling argument for me. You think Google has had some PR nightmares before? This kind of outright betrayal of trust would be devastating! and for what, only a minute sliver of an edge.
I work with a whole lot of domains. Some of them have added GA after I was privy to their private data. There's never been a hint of that addition causing a problem. Not from adding WMT, GA, CSE, Adwords, Adsense, YPN, Sitemaps, AdCenter, or any other of the theories I see floated around the web.
None of these ideas holds any water - not even from a Google business perspective, to say nothing of the data I look at every day.
When we think about Google, we often project onto them the way we think and operate and leverage things at our own scale of operation. And then there's the impulse to expose something scandalous - the way the big media do every day to keep their ratings up. None of that will improve a website, and that's where a webmaster should be focused.
So I don't think it's unreasonable to draw a link (though with talk of known penalties, it becomes a bit of side issue!).
Personally, I'd take it as a friendy piece of advice to fix the site, rather than remove the code (too late), or 'not use it' in the first place. It seems to me that using Google's tools to warn of Google's displeasure is wiser than waiting for the Heavens to fall in. But YMMV.
All our sites used to get heavy traffic from google..
Are "all" these sites by any chance interlinked? If so this could be part of your problem, i.e., google may consider your sites a network set up for the purpose of increasing your PR.
This sites are not interlined with each other. 3 sites reside on server A and rest all sites have there own servers with different hosting providers (ofcourse different class c ip)
Sites widely affected are forum based site(s) like webmasterworld where traffic suddenly dropped from 95% to less then 2% from google all alone..
Happened with GA immediately in 3 days and with CS in around a month .
There are no changes made to server between this period neither any software upgradations, although we added more RAM to same.
When you signed up for GA and CS for all of the affected websites:
- Did you give Google the exact same name, contact info, etc for each site?
- Did you select "anonymity" for each of them? (re: Marcia's post above)
.......................................
