If you are an Analog CE user, you may have noticed that the Search Query and Search Word reports suggest that your site has been receiving little to no search engine traffic. The Search reports may be empty or sparsely populated with results. This article discusses the cause of the problem.
What are the External Search Reports?
The premise of the external (rather than internal) search report is to show you what a visitor typed into a search engine before landing on your site. The Search Query report shows the entire search phrase used, while the Search Word report creates a word count report for anything de-marked by a space.
They are intended to help you identify what your users are searching for as part of your SEO activities. Unfortunately, over the last couple of years the data available on the report has been declining. On some sites, the report may have disappeared altogether as no data is available to Analog CE.
Fixing the empty or sparsely populated report
There are three reasons for the decline in report quality. They are outlined below in order of lest to most severe.
Missing Search Engines
The default config file in the Analog CE source release is a clone of the original Analog 6 config file. It (and your own) config files have not been updated with modern search engines. Consequently traffic from Bing and DuckDuck Go (and others) will be wholly absent.
I will be including an updated config file sample in Analog CE starting with Analog CE 6.0.16. Analog CE users will need to replace or merge this into your running configuration files. To perform this manually, you must add relevant
SEARCHENGINE entries to your global/site config files.
HTTP > HTTPS
As a result of security issues and the gross mismanagement of private data transiting public networks by the CIA. The world made a rapid switch from HTTP to HTTPS in 2015. The default Analog 6 config file only included Search Engine detection definitions for unencrypted HTTP connections. Consequently Analog CE has stopped reporting traffic from any HTTPS search engine source.
Starting with Analog CE 6.0.16, I will fix this in the bundled sample file. You will need to merge the
SEARCHENGINE changes into you respective config files.
Privacy & Commercialisation of Analytics
The crux of the problem stems from Search Engines no-longer send the search query in their referrer header. Some see this as being for legitimate privacy reasons, others for commercial benefit. DuckDuck Go would certainly argue the former. For Google, the reason is less clear. With up to 98% market share, Google is effectively forcing webmasters to use Google Search Console and/or Google Analytics to understand their user needs. In turn, this feeds Google with more accurate data on user activity for ad-profiling. It also ensures that you as a webmaster are creating the web that Google wants, above any other concern. Any webmaster threatened with de-listing due to mobile device incompatibility will understand this.
The following sample log line is from the 5th April 2015
2015-04-05 15:38:29 W3SVC5 web02 xxx.xxx.xxx.xxx GET /downloads/msie/ie60sp1/ - 80 - yyy.yyy.yyy.yyy HTTP/1.1 Mozilla/4.0+(compatible;+MSIE+4.01;+Windows+98) - http://www.google.nl/search?q=internet+explorer+download+windows+98&ie=ISO-8859-1&hl=nl&source=hp&gbv=1 www.hpcfactor.com 200 0 0 10747 331 406
This is a similar one from 4th June 2019
2019-06-04 00:17:31 W3SVC5 web01 xxx.xxx.xxx.xxx GET /hardware/devices/specification.asp d=142 443 - yyy.yyy.yyy.yyyy HTTP/2 Mozilla/5.0+(Linux;+Android+6.0;+HM-G552-FL+Build/MRA58K)+AppleWebKit/537.36+(KHTML,+like+Gecko)+Chrome/56.0.2924.87+Mobile+Safari/537.36 - https://www.google.com/ www.hpcfactor.com 200 0 0 12534 491 495
If you scroll to the right, you will notice that in the earlier file, the user searched for “Internet Explorer Download Windows 98”. In the latter file, Google only informs the server that the request came from google.com. The context of the search is no longer offered.
As a consequence, the data required to create the Search Query report is not available to Analog CE. This is the reason why the Analog Search Query Report is empty or sparsely populated. It is not a bug. Rightly or wrongly, it is a symptom of the modern web.