Showing posts with label amazon. Show all posts
Showing posts with label amazon. Show all posts

9/26/2014

Google Analytics: How to filter out "Amazon Technologies / Boardman" bot that creates inaccurate stats

My work server has been getting clobbered lately by some bot program at Amazon.com.

As a result, our Google Analytics visitor and Pageviews are off by thousands each month.
While it might look good in the numbers, it is terribly inacurate as far as real-world visitors are concerned. 

How to Check for Amazon's Awful Bot

Data proving this can be found in Analytics in a number of ways:

Check Method #1

Analytic Reporting > Audience >  Geo > Location > [ Explorer ] tab
     > United States > Oregon > City 

If you see a ton of visits from "Boardman," then your server is getting inundated by Amazon's servers, too.

Check Method #2

Analytic Reporting > Audience >  Technology > Network  

You will likely find, under "Service Provider," two listings for:
  • amazon.com inc.
  • amazon technologies inc.

Check Method #3

Analytics Reporting > Behavior > Site Content > All Pages
     > [ Seconary Dimension] button > Network Domain

You will likely see amazonaws.com 

How to Filter out Amazon's Run-Amok Bot

So, now that you are sure that some bot on Amazon's server is messing with your stats, you can create a Google Analytics filter.

Filtering is not as obvious as it should be, but creating one is not hard.

First, you need to be logged in as a User with the ability to Edit.

Next, go to the [ Admin ] tab at the top.
Make sure your domain is selected.
Click on  [ All Filters ] tab on the left.
Click on [ + NEW FILTER ] button.




Now, you're going to make two filters,
If you happen to live in Oregon or need to include the city of "Boardman," then perhaps only the first one.

Filter #1:

Filter Name: "AmazonAWS (bots)"
Filter type: Custom filter
Choose: Exclude
Filter Field: "ISP Organization"
Filter Pattern (very specific & without quotes): "amazon technologies|amazon\.com"
Case Sensitive: No
[ Add >> ] YourDomainName.com
Click [ Save ].

Filter #2: 

Filter Name: "Boardman (bots)"
Filter type: Custom filter
Choose: Exclude
Filter Field: "City"
Filter Pattern (very specific & without quotes): "Boardman"
Case Sensitive: No
[ Add >> ] YourDomainName.com
Click [ Save ].


That should take effect in a few hours.

The existing data will not be affected retroactively -- meaning, the Amazon / Boardman filters will only be applied to data from this point on.

You can check to see if it's working by doing the checks above in two days.
If the counts for the bots have dropped significantly or gone to zero, then you know your stats are being filtered and should be back on track ... at least for a little while.

Comment below and let us know if this worked for your.