Google Encrypted Search Strips Referrers, Hides Visit Source

Web analytics tools loose the HTTP referrer of new visitors that come from Google’s encrypted web search. The issue arises when the browser moves from an encrypted URL (https://) to an non-encrypted URL (http://) – which are most of the web sites on the net.

The effect on web analytics tools is that visits will appear to be coming “direct” when in fact they are originating from search.

Stay tuned.

AVG Anti-virus Software Generates fake traffic

Just a heads up that anti-virus maker AVG has included a new feature in it’s AVG 8 software that generates fake visits to your web site.

The way this works is someone running AVG 8 (there are 20+ million people) searches for something using Google. AVG will, in the background, visit each web site listed in the Google’s results in order to check and see if the web site might contain malware or virus type stuff.

A nobel idea, but a nightmare for web analytics because AVG is disguising itself as a normal browser when doing these checks. Instead of properly identifying themselves, AVG uses the user-agent of IE6 – essentially making itself indistinguishable from a real web browser.

This should not effect OWA due to the super-secret checks that OWA does to see if visitors are really robots.

Stay tuned for more on the state of user-agents and inflated traffic numbers shortly.


Is Web Analytics Dead

Anil blogs his response to the question: “Is Web Analytics Dead?” and concludes that no, it’s just maturing into a holistic platform for gaining a 360 degree view of the customer.

Personally I think companies are still a long way from having that 360 degree view of their customers due to the fact that web analytics data is mostly stored in a disonnected or outsourced silos and systems.

There are many reasons for this, but I blame the lack of strong open source web analytics efforts like the ones that exist around other areas of web application development (i.e. web servers, or content management, or databases, or tagging, etc.).

Think about it. When a developer sets out to build a new web application for their company, they don’t spend any time writing code to serve the web pages. Instead they use an open source web server like Apache and rely on its features. In fact, frameworks like Ruby On rails or the Zend framework now allow the developer to bypass spending time on a whole lot of lower level application code and base services like authentication and database access.

However, when it comes to web analytics, developers have historically been out of luck when they reached for the an open source solution to the problem. This has put their company on a collision course with outsourcing their web analytics and data management to a 3rd party.

Working with a 3rd party service provider allows a company to put a lot of base capabilities in place quickly, but can begin to cause problems when it comes time to link web analytics data to detailed customer records stored in internal databases.

Right now that is hard to do for most companies that have outsourced their web analytics to a 3rd party service provider because that service provider either does not provide access to the raw event level data, or they set their own tracking cookie with their own unique visitor ids.

We are hoping to reverse this trend by providing developers and companies with a full featured toolkit that they can use to quickly add web analytics capabilities to their own web applications.

Our goal is to ultimately provide plugin level integration for OWA with all the major application frameworks that developers are using today. For example, if you are developing an PHP based web application or using WordPress or MediaWiki then you can use OWA to add web analytics functionality to your application in minutes.

If this sounds familiar to you or your situation, then you might want to give OWA a try.