Does Removing Google Analytics Help Your Conversion Rates?

[samwest]

During the usual zombie talk and between August SERP changes discussions, a few members who were long time lurkers popped their heads up and started suggesting that everyone with zombie traffic try removing Google Analytics tracking code and removing Adsense Ads, your webmaster account and everything Google. You are also advise to block moz and ahrefs in you .htaccess. Seems like a bad joke, but they argue that Google is using our own traffic data to plot evil schemes to load their pockets. So, what's your take on what even I consider to be intentional misinformation or Google paranoia. Hopefully this post will offload the discussion to a side topic rather than the monthly SERP changes hangout. Feel free to share.

[Robert Charlton]

Mod's note: The submitted title and description line for this thread were...

Does Removing Google Analytics Help Your Site?
Is Google using your visitor data against you?

Since we don't use the description line in this forum, and the question wasn't asked in the original post, I'm posting it here.

Additionally, the question of conversion rates seemed to be what was distracting the Updates Thread [webmasterworld.com...] ...so I'm adjusting the title here to reflect that particular concern.

I rely on thread participants to fill in further.

[AussieWebmaster]

Would love to see someone test this lol but what would it be converting

[Nutterum]

I'd like to follow up on this thread. I can't remove GA from sites that actually are eligible for this test, but I am curious to see what people see.

[glakes]

There's no disputing that Google's profits have risen sharply at a time when our results from Adwords were at their worst. I use the term "were" because to stop the bleeding and negative ROI, my Adwords campaigns remain off most of the time except for testing. So any conspiracy theories have their merits IMO. All we have to do is use our common sense to follow the money.

RankBrain seems to be the most plausible and non-evil explanation for why we are seeing zombies. Where does RankBrain get its information from? Possibly our Analytics accounts? If so, I'm wondering if removing GA would force RankBrain to eventually reboot and reassess or disregard our websites. If this were the case, GA may have to remain off our sites for some length of time to extend beyond whatever data freshness cycle RankBrain uses. And this would have to coincide with RankBrain being taken offline to learn when the data it has on our sites is no longer fresh and reliable - possibly taking our sites out of the mix when it comes to rearranging our sites in queries to match what RankBrain perceives users are looking for.

Having the zombie problem for 11 months so far, it is also possible that RankBrain has too much to learn before it reassess our high bounce, low page view and poor conversion rates. Once again, digesting data from GA would reveal the blatant truth that the zombie traffic Google is sending in both free and paid is rubbish. Though managing searcher expectations and profitability may be a primary function of RankBrain, and our GA accounts would be a major asset in Google maintaining such a balance.

[samwest]

I'm thinking this has more to do with the heavy value placed on back links. A site with 1.2 million back links will be able to post a one page article that dominates the serps while a comprehensive detailed site with hundreds of pages of good / useful content with 1.2k back links will suffer forever to rank at all. What really sucks is to have survived and lived through the successful earlier days when back links were not as valued...then to watch the slow demise. I'm not sure removing anything at this point will help...and like glakes said, it may take a very long ting for their black box to reset on your site...if ever.

[edited by: samwest at 1:07 pm (utc) on Aug 18, 2016]

[aristotle]

Analytics data can't be of any use for determining search rankings and traffic because a lot of sites don't have it. As an example, I've never had Analytics on any of my sites. If a lot of sites don't have it, the data is incomplete and therefore can't be used for comparing different sites with each other.

In addition, If turning Analytics on or off had any effect on a site's rankings or traffic, this would have been discovered years ago and would be well known in the SEO community.

The Chrome browser can collect data from all sites, and it is much more detailed data on user behavior than Analytics can collect. In fact the main reason that Google expended all the money and effort to create the Chrome browser was to be able to collect this data.

[samwest]

^ exactly - so I think we can safely assume that the advice from Dooku is doodie.

[bakedjake]

Does Removing Google Analytics Help Your Conversion Rates?

No.

[glakes]

I'm thinking this has more to do with the heavy value placed on back links. A site with 1.2 million back links will be able to post a one page article that dominates the serps while a comprehensive detailed site with hundreds of pages of good / useful content with 1.2k back links will suffer forever to rank at all.

My site is very clean and has few links to it and ranks fine for buyer keywords when I check it. The rank does change a little when doing a search from different IP addresses but nothing major. What ranks good for me may not even be seen by a buyer if RankBrain is re-ordering search results on the fly. With that in mind are you saying to get buyer traffic one needs a lot of links?

@aristotle

Analytics data can't be of any use for determining search rankings and traffic because a lot of sites don't have it.

Just because Chrome can collect data does not mean it is. Chrome does not have full marketshare and phone home features can be disabled.

The Chrome browser can collect data from all sites, and it is much more detailed data on user behavior than Analytics can collect.

Chrome has a 59.5% marketshare - Source: [w3counter.com...]
Google Analytics is installed on 8.4% of all websites and 54.6% of the top million websites - Source: [marketingland.com...]

The argument that Chrome will collect all data implies it is being used by enough people to accomplish the task on its own, and its not. On my site 37% of users in the last month were using Chrome. Removing one of these components (Chrome or GA) from RankBrain's interpretation model/equation may leave too much open to determination and lead to unreliable results. Therefore it may take both Chrome and GA to give RankBrain a more complete picture of how to rank ones site.

[aristotle]

Chrome has a 59.5% marketshare

That just verifies what I said - that Chrome can collect data from ALL sites, or virtually all sites, the exceptions being sites that either block Chrome or almost never get any traffic at all.

Also, Chrome collect far more detailed data on user behavior than Analytics can.

It was because Google realized the deficiences of Analytics that they decided to expend all the money and effort to create chrome.

[blend27]

I had it on one of my sites in 2004, removed it and never had GA afterwards or recommended to any of my clients to have GA on their sites.

Still all of the sites get Zombie traffic. ECom sites get more that others in many of my cases, that is a Fact.

Plus with the rise of Ad-blockers and Tracking protection, what good would it do anyway. You don't get a complete picture, they(goog) don't get a complete picture either.

But if one wants to remove it and do some A/B tests - God bless...

[glakes]

That just verifies what I said - that Chrome can collect data from ALL sites, or virtually all sites, the exceptions being sites that either block Chrome or almost never get any traffic at all.

You ignored my point. Chrome is not collecting data from all sites because it does not have enough reach on its own and phone home features can be disabled. Only 37% of my visitors in the last month used Chrome. I would think that any machine learning process would require more than 37% of the facts to make an informed decision. Couple Chrome with GA and the data pool grows, which may allow RankBrain to make a more informed decision.

Whether Chrome can collect data and send it back to Google is not in dispute. Chrome is one of the many ways Google tracks and profiles users. What is in dispute is whether our individual websites have enough Chrome traffic on its own to feed RankBrain enough information. In my case, the answer would be no. GA may be filling the gap in data.

[mosxu]

We actually did remove GA a few times and I would say that it helps especially if competition is high as it makes it easy to measure engagement but also can be a problem if fake users are trying to influence CTRs.

But GA or Piwik cannot be considered when it comes to understanding traffic and optimize sites for conversions. So we have turned to clicktale, there are some others but do not recommend hotjar as it sampling the data. So easy to watch a video of every visitor journey. It changes your mind about the quality of the traffic some sites receive.

[Shaddows]

To clarify what others are saying, because some people are talking at cross purposes...

With Chrome, Google knows what chrome users do on 100% of websites
With GA, Google knows what 85% of users do on one specific site

Google then extrapolates across a large number of GA sites that
"When Chrome users do THIS, then all other users do THAT" (Which will obviously have a large intersection on your Venn diagram)

Thus GA can be used to learn typical differences between chrome/not-chrome, then use that data to approximate all interactions on non-GA sites using just chrome users.

That's the theory.

Now, I don't think it's controversial that Google uses behavioural data of some form to influence SERPs. This user behaviour, possibly but not definitely.

However, given that there are enough tracking-enabled chrome browsers on your site, trained against sites "just like yours", removing GA will not make much difference whether they use it or not.

[aristotle]

glakes -- You need to get a better understanding of the fact that there are hundreds of millions of Chrome users, and that over time some of them will visit almost every site on the web.

[glakes]

You need to get a better understanding of the fact that there are hundreds of millions of Chrome users, and that over time some of them will visit almost every site on the web.

I understand that 37% of my visitors in the last month used Chrome, well below half. Also, you need to quit using Chrome to paint with a broad brush. There are two major flaws in your theory as it relates to my site.

1. Only 37% of users used Chrome in the last month of which some could have disabled call homes either through Chrome or firewall.
2. The biggest flaw in your theory yet - My largest customers (biggest spenders) are other businesses and do not use Chrome when ordering from work. Often these people are restricted to whatever the network admin installed on their workstations, which is IE (of all things) for my largest customer. Meaning that if Google relied solely on Chrome for all of its data, it would not have any data on the users that spend the most on my site.

With that said, I think collecting data from Chrome may work for Google on some sites but on others would produce a picture so inaccurate it would not be useful. And that's where GA comes in to collect data from a source that Chrome does not have access to.

[aristotle]

The biggest flaw in your theory

I think you may be talking about a flaw in google's algorithm, not a flaw in my theory.

Anyway I don't understand why there would be this much confusion about this matter. It's really not that hard to understand.

[londrum]

if your bounce rate is bad, and your visitor time on site is bad, then i can maybe understand why you wouldn't want google to know (even though they'd inevitably find out anyway)
but if your stats are decent, then what's the harm in having analytics? all you're doing is making sure they're aware that people are using your site, instead of leaving it up to chrome users.

the only other reasons for taking it off are if it slows your site down, if the information it provides is of no use, or if you've already got another stats package going

[ecommerceprofit]

if your stats are decent, then what's the harm in having analytics?

"IF" Google did not like that you are running ads that are profitability working (low conversion cost) then it would be in their interest to raise your conversion costs to make more money. Not saying they are doing this but I think it would be quite easy for them to take advantage undetected.

No one has access to their system from the outside to verify one way or the other. I would think that being the biggest or one of the biggest lobbyists in the country stops regulation from the FTC. This lobby theory is only a theory and should not be taken as fact.

[glakes]

I think you may be talking about a flaw in google's algorithm, not a flaw in my theory.

You are the one that brought up Chrome can collect data from all sites and that removing GA would not matter, which is a theory I don't agree with.

Anyway I don't understand why there would be this much confusion about this matter. It's really not that hard to understand.

It is very easy to understand. I pointed out a real life example where Chrome can't collect data by some of the largest spenders on my site because these shoppers are employees behind a business network that does not use Chrome for a web browser. Therefore, in the context of this thread's topic that removing GA can't possibly improve conversions because Chrome collects the data anyway is not accurate. When orders come in for $x,xxx from a user with IE (repeat customers and direct visits), without GA Google would not know the transaction even occurred.

Now, let's assume that Google is only using Chrome to collect data to feed RankBrain since the theory is out there. In doing so then RankBrain would not have any information related to those customers that have the largest orders, even though they are reported in GA with ecommerce tracking. Could this be the reason for zombie traffic and poor conversions?

[NickMNS]

To further the point made by Shaddows above, the other issue is that since the number of sites using GA is large and constant (in the order of magnitude). Whether one, two or even 100 sites decide stop using GA, this will have no impact at all on the performance of the algorithm, if one even exists. So from that perspective alone this whole theory of removing GA makes no sense.

In the Serps thread that spawned this Wijnand schouten, made the argument that removing GA would prevent competitors from remarketing to users of your site. Again, this is misinformed. remarketing is accomplished by dropping cookies on the users computer when the users first visits the competitor's website, when the user then navigates to your site Adsense detects the cookies and serves the ads for your competitor. No GA required, GA was mentioned in the links he posted to allow you to track what your users are doing when you are using remarketing. The solution to prevent competitors from advertising on your website is by simply blocking their URL in Adsense.

One other big problem with conversion tracking in GA is that it is completely subjective and unverifiable (from Google's perspective), you can make almost any action a conversion and assign any value to it that you would like. Google has know way of knowing whether what you input is accurate or not. Google cannot calculate your ROI. In fact the theory can be tested, set the conversion value very high in GA and see if this has an impact on your Adwords cost.

A more plausible scenario as to why adwords cost rise to point were you are no longer profitable, is simple economics. If you sell a widget at 10$ and your COGS (cost of goods sold) is 2$ and adwords cost is 1$, then you have margin of 7$, Yeah! A new competitor enters the market, sees all the profit to be made (7$ widget) so he decides to sells almost the same widget, but in order to gain market share he simply doubles what you spend on adwords and so he earns a 6$ margin. But of course before he eats your lunch, you jump in and spend 3$ and on and on, until, your gross margins just barely covers your fixed costs, and since your costs are similar to your competitors you both end up paying the max you can afford on adwords, you split the market 50/50 and earn meager profit margin, all while Google is sitting back counting the cash. No conspiracy, no dubious practices, simple economics.

[aristotle]

glakes -- I still don't understand what you're trying to say. I'll just point out that I've never had Analytics on any of my sites, yet somehow Google has been able to rank them anyway.

[glakes]

@ aristotle

My comments are in response to the subject of the thread - does removing GA improve conversions. I'm not saying it does or does not. What I am saying is that since 9/2015 my conversions from Google took a huge tumble, which was right around the same time as many others. I, like many others, want to know why conversions from Google took such a huge hit. My ranks are just fine for buyer keywords when I check and my ranks really have not changed in a very long time. The only change I have seen is in the quality of traffic from Google. I'm being nice by focusing on RankBrain as a possible reason for the traffic quality plunge, but in reality we all know Google is out to make a buck as we all are.

Those of us with zombie problems saw on 8/1 a major relief in zombies. For most of that day traffic from Google was converting like it was prior to 9/15, then wham it all stopped. I'm assuming that it is likely RankBrain was offline for part of its learning lessons, but why 8/1 changed is anyone's guess. But we do know RankBrain is taken offline to learn if what Google previously said was/is still true. RankBrain has to get its data from somewhere beyond search. Chrome, as you noted, is a good source of data for RankBrain. So is GA as it catches some of the user data that slips through the cracks from users not using Chrome. There are other data collections too, but that is outside the scope of the discussion we are having.

The bottom line is I want to rid my site of zombie traffic from Google. From a business perspective it only makes sense to try to correct what is wrong (if anything) and reap the benefits of increased sales. If it is RankBrain that is to blame, then either the data it has is incomplete or inaccurate in some way. If we were penalized, we would not see much traffic at all. We are seeing a lot of Google traffic, even higher then what we were getting before the zombies hit. If RankBrain has bad data or is downgrading or removing our SERP positions on the fly to buyers what can we do to force it to reboot, reassess or relearn? Would removing data collection points, such as GA do the trick? Would sending all Chrome users originating from Google Search a 401 help? I'm serious, the traffic Google is sending is absolutely junk.

@NickMNS

There are no new competitors in my industry. Organic ranks are great when I check them. Bidding Adwords category recommended bids, with an ECPC of even 50% did not change the poor conversions. Today for example, Google sent four times the traffic as Bing and less then half the conversions. This is not a bidding problem or an organic rank problem. It is a Google traffic quality problem.

[smilie]

--
@aristotle: Analytics data can't be of any use for determining search rankings and traffic because a lot of sites don't have it. As an example, I've never had Analytics on any of my sites.The Chrome browser can collect data from all sites, and it is much more detailed data
--

You missed the topic of a thread. The question is not where to collect better data about websites.
The quesiton is "does Google use Analytics AGAINST WEBSITE OWNERS , at all or on some occasions".

The definite "YES" answer to this happened when Google bumped everyone's Adwords Pay Per Click costs from 10 cents
to over $1 and more. In this particular case, it DID NOT MATTER that you specifically had or did not have Analytics installed,
Google went by the data collected from your competitors.

[smilie]

@bakedjake : No

Have you actually tried on a conversion-specific site, such as small-to-medium ecommerce? And long enough.

[Simon_H]

Paraphrasing what I said on the other thread... I think it's silly to suggest that Google uses GA in this way. For Google to use GA data to control converting traffic, it would require conversion tracking to be used, which is rare and never implemented consistently. Same point as @NickMNS above.

But if you really want to test the theory, turning off GA for a few days won't prove much. Instead, firstly, change your payment complete page to only include the GA completed conversion code for 1 in 10 transactions, i.e. trick GA into thinking you have far fewer sales than you do. Leave like this for a few weeks at least. Then change your site to show the GA completed conversion code too much, e.g. add it to the billing page instead of the payment complete page. Leave like this for a few weeks at least. If the theory is correct, you should see your real sales or traffic change between the two as Google compensates for the (fake) increase in quality users. But, frankly, you'd be wasting your time because I'm fairly sure there would be no change.

[bakedjake]

Have you actually tried on a conversion-specific site, such as small-to-medium ecommerce? And long enough.

Nope. I don't need to remove what is probably the single most used piece of JS in the world (not to mention a critical business system!) to know that doing so won't help me sell more product. The claim is absurd. The onus is on those making it to prove it.

I'm more than happy to help you try it on your site though!

[edited by: bakedjake at 5:13 pm (utc) on Aug 19, 2016]

[smilie]

>> @NickMNS: If you sell a widget at 10$ and your COGS (cost of goods sold) is 2$ and adwords cost is 1$, then you have margin of 7$,

This is what Google thinks too. It is a completely flawed logic.

If your conversion is 1%, even if it is above average 2% for ecommerce, you are out of business, because you need to spend $50 to get one $10 conversion.

Not only they don't pay attention to this in their analytics big data rollup to bump Adwords prices and call it "market", they also completely ignore returns, my cost of doing business, and you name it 20 other things.

[bakedjake]

Still all of the sites get Zombie traffic. ECom sites get more that others in many of my cases, that is a Fact.

Would you be willing to publish some of your logs to analyze?