Web Analytics Data driven business freaks me out
I used to be a web developer. A deviant one. I had no computer security knowledge, but I have found several flaws. They were obvious and careless ones. Because of this, most of the time, my manager did not believe me. But, I think most people have too much trust in editors and the open source community to believe that they can discover those types of errors.
Today, I am not a web developer anymore. I am a web analytics consultant. I help my customers improve their websites using web analytics tools. But, I am still mistrustful.
The trend is to have faith in data. Data seems better than intuition and experience because it is supposedly indubitable. But the thing is: it is.
Web analytics tools are world writable. They are based on small transparent images called beacons that are automatically retrieved using javascript. But, it is very easy to manually write fake orders on any analytics solutions by running a few lines of Javascript or forging HTTP requests to this beacon.
This is a well-known issue. A large number of people are aware of it but still use analytics data to:
- test variants of their website
- pilot paid traffic campaigns…
- sort items depending on their popularity
- recommend items depending on the user behavior
- send abandonned cart emails
Do you think it is really sane to only rely on world writable data for this?
Your competitors may forge fake data to push you to make bad investments and decrease the quality of your website.
Unfortunately, this problem is easier said than done . It is not that easy to solve. In my opinion, web analytics tools will never provide 100% reliable data .
To mitigate the risks, I recommend always using several sources of data (e-commerce databases, affiliation platforms, different web analytics data sources, a/b testing solution reports…). The metrics provided by these different sources may be slightly different. It is understable because the metrics have different definitions and are not processed the same way. Thus, you have to check that these sources of data are consistent.
I do not pretend web analytics is dangerous and shoud not be used. I just mean data should not replace instinct, critical mind and experience.
Keep in mind that you can never trust web analytics data 100%.
