Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
Well this looks like the most suited place for this question.
Every website collect data of the user, some just for usability and personalization, but the majority like social networks track every move on the web, some free apps on your phone scan text messages, call history and so on.
All this data siphoning is just for selling your profile for advertisers?
A couple of days ago developers from one product company asked me how they can understand why new users were leaving their website. My first question to them was what these users' profiles looked like and how they were different from those who stayed.
Advertising is only top of an iceberg. User profiles (either filled by users themselves or computed from users' behaviour) hold information about:
So it's more about improving product and making better user experience rather than selling this data to advertisers.
Most companies won't sell the data, not on any small scale anyways. Most will use it internally.
User tracking data is important for understanding a lot of things. There's basic A/B testing where you provide different experiences to see which is more effective. There is understanding how your UI is utilized. Categorizing your end users in different ways for a variety of reasons. Figuring out where your end user base is, and within that group where the end users that matter are. Correlating user experiences with social network updates. Figuring out what will draw people to your product and what drives them away. The list of potential for data mining and analysis projects could go on for days.
Data storage is cheap. If you track everything out of the gate, you can figure out what you want to do with that data later.
Scanning text messages is sketchy territory when there isn't a good reason for it. Even when there is a good reason it's sketchy territory. I'd love to say that nobody does it, but there have been instances where big companies have done it and there are a lot of cases where no-name apps at least require access to that kind of data for installation. I generally frown on that kind of thing myself as a consumer, but the data analyst in me would love to see if I could build anything useful from a set of information like that.
Here's a practical example of using web data for something other than advertising. Distil Networks (disclaimer, I work there) uses network traffic to determine whether page accesses are from humans or bots - scrapers, click fraud, form spam, etc.
Another example is some of the work that Webtrends is doing. They allow site users to build a model for each visitor to predict whether they'll leave, buy, add to cart, etc. Then based on the probability of each action you can change the users experience (e.g. if they're about to leave, give them a coupon). Here's the slides from a talk by them.
