3

Everyday :(, the following user creates one hit on my google analytics which has 3-5 sessions and %100 bounce rate.

ISP: Digital Ocean

Operating System: Linux

City: san francisco

I want to block this, but can't find the IP.

SOLUTIONS I TRIED:

1- I installed a plugin (WP Statistics) to track user IPs, but it can't catch this user (or bot or whatever)

2- I recorded every IP address that hit my homepage, using following php code piece, but I can't catch the user.

$_SERVER['REMOTE_ADDR'] 

I will be happy to hear your suggestions.

HOY
  • 504
  • 1
  • 5
  • 17
  • 3
    Is it worth blocking one IP even it may generate 100% bounce rate? There would be many situations like this, blocking is not the solution unless it might be causing issue which may not be i guess. – TopQnA Jul 24 '17 at 00:40
  • 1
    It is likely not someone who hits your sever but rather is using your GA code to spam or create some sort of havoc in GA. – closetnoc Jul 24 '17 at 01:17
  • @TopQnA , I believe so, there are 2 reasons. The first one is one by one the number of spams may increase, and it will be harder to detect them when my site gets more traffic, so I have to be in control from the begging. So it is actually not block just 1 IP. The second reason is that my site is a small site, and I am tring to improve my ranking on google and with high bounce rates it is harder. I believe that in 2017 bounce rate, and time on page is important factors to rank on google. Here is a video about it. https://www.youtube.com/watch?v=lyiikzjg9a0 – HOY Jul 24 '17 at 08:54
  • 3
    I am aware of bounce rate and the impact on ranking etc. You also need to understand that traffic from the same IP (One or few) with higher bounce rate would have almost no impact. So, in my view, it will be just an overhead try to deal with it. Instead, same time could be used to improve the website quality etc. Google is quite aware of basic issues in general. I would say be productive. – TopQnA Jul 24 '17 at 08:59
  • 1
    @HOY Google Analytics data has no bearing on Google search ranking. – DisgruntledGoat Jul 27 '17 at 12:02

1 Answers1

6

This problem is related to spam actions against your GA account. Neither the WP plugin nor the php code would be able to record ghost visitor's (aka Ghost referral) IP address because they are hitting your GA account making HTTP requests by sending raw user interaction data directly to Google Analytics servers. In other words, they do not actually "visit" your website.

The best you can do is filtering these bots via GA Filters. Follow this post recommendations or this other one. I particularly like filtering by "hostname".

Hope that helps.

Raul Reyes
  • 2,476
  • 12
  • 22
  • It is a crying shame that after all these years Google has not solved this problem and does not appear to want to either. – closetnoc Jul 24 '17 at 01:18
  • @closetnoc it is a crying shame indeed. – Raul Reyes Jul 24 '17 at 01:34
  • I'm not sure they could completely solve this issue. I think the best they could do is make it more difficult for the spammers. – Stephen Ostermiller Jul 24 '17 at 02:48
  • @StephenOstermiller It would be nice if they tried. I would not be critical if they at least pretend to give a damn. After all, it is their product! There is a simple solution available that would work with the existing system, however, each site would have to change their GA id. The solution existed since day one. – closetnoc Jul 24 '17 at 03:45
  • @closetnoc Exactly how do you propose to reliably detect this problem? If you have an answer that can actually detect the difference between failed and malignant connections (for example) with zero false positives, I'm sure that Google (and a lot of other companies) would be willing to pay you - probably a lot of money - for your solution. – FKEinternet Jul 24 '17 at 06:07
  • @titico I believe filtering only removes data from the user interface, that is why I have concerns, in the documentation you sent, there is a video which says google analytics data is not affect ranking seo, but I don't think this is true in 2017 according to following seo video. https://www.youtube.com/watch?v=lyiikzjg9a0 (Which is the reason why I am tring to reduce my bounce rate here) – HOY Jul 24 '17 at 08:46
  • 2
    Google doesn't use GA data for ranking purposes. For SEO they use a similar metric: Bounce-Back-Rate (BBR). BBR is measured by Google independently of GA. See https://webmasters.stackexchange.com/a/54336/ for my explanation. – Stephen Ostermiller Jul 24 '17 at 10:09
  • @FKEinternet Being snide? A solution has existed either on SE or another well known site. It is a simple change where the ID cannot be guessed or read. Simple pimple. – closetnoc Jul 24 '17 at 18:46
  • @HOY Algorithms require consistent and reliable data. Since most sites on the web do not use GA, any metrics from GA cannot be used. Google search only uses what data the search engine itself can collect. Not data from opt-in offerings. You are okay. – closetnoc Jul 24 '17 at 18:50
  • @closetnoc Not at all. I'm not sure what the "solution" you mention is about, but an "ID" has nothing to do with how the HTTP protocol works - if someone opens a connection to a server with telnet example.com 80 and immediately disconnects, how do you differentiate that from someone trying to access example.com with their browser and their system crashing (which causes a disconnect)? (Note that there's no "ID" involved here.) – FKEinternet Jul 24 '17 at 19:17
  • @FKEinternet The problem is two fold, one the ID can easily be read from the website being effected, or two the ID can be easily predicted. Worse yet, reasonablely predicted IDs can easily be generated randomly and used to spam GA. The solution was to issue new unpredictable IDs and change the GA code not to require the ID within the JS. Within days what appears to be a workable solution was offered. However, with filtering as a work-around, Google does not feel that anything more needs to be done. This has gone on long enough that G should have cracked this nut by now. Cheers!! – closetnoc Jul 24 '17 at 20:02
  • @closetnoc What is this "ID" that you're referring to? I don't see anything in the question or this answer that has anything to do with an "ID" – FKEinternet Jul 25 '17 at 04:14
  • I have not looked at GA code in about 2 years, however, Google would issue a unique identifier to insert into the JS code for the site. – closetnoc Jul 25 '17 at 04:29