Definition
The process of removing personally identifiable information from data sets, so that the people whom the data describe remain anonymous. – Wikipedia
Relevance in CRO
Online privacy is a growing concern. Users don’t really want every aspect of their lives recorded, analyzed, and crunched through algorithms so companies can shove targeted ads in their faces. Laws like GDPR and CCPA and are being passed to protect individuals’ privacy and control over their own data. Browsers are locking down and potentially eliminating cookies. Apple is stripping tracking pixels from emails.
I guess we should pack it up then; if you can’t track user behavior then how can you possibly run A/B tests? Well… that’s not quite true. Online experimentation is about determining how a change affects the behavior of users (plural) not the user (singular). It’s all about behavioral data, not identification data. An experiment is evaluated based on what a user does, not who they are. You’re interpreting the total result. Any personally identifiable information either isn’t collected or is stripped away. OK sure, with a wide enough data set you can probably do fingerprinting, but a compliant system will prevent that through prevention, aggregation, and anonymization. These processes are what allow A/B testing to still be possible in a privacy-compliant fashion.
Additional Reading
- GDPR -GDPR.eu
- CCPA – California’s Office of the Attorney General
- The differences between anonymous and personally identifiable data – Kameleoon
- AB Testing Application Compliance – Convert