Subscribe by Email

Sunday, April 28, 2013

Capturing information from within the product - Need to balance the privacy perspective

This is a post that is again based on real life experience. Let me lay out an example to you. Lisa is the program manager of a team which is developing a new version of a software that prepared greeting cards. The software allows the users to add their own images or use images that are available within the software, and you can do the same for greetings and videos, and the output is a rich electronic greeting card that can be sent out by email or posted on social networking platforms, and it even provides a more plainer output that can be printed.
Now, this card needs to work on many different operating systems, needs to work at machines with different performance parameters, and also needs to work on different browsers. How do you get this sort of information ? Well, you can get this information from users in the terms of mailers, surveys, questions from within the software, and other such ways, or you can get it from the users own machines. So, Lisa did some discussion with her team, and came out with a technical solution - there will be some tracking devices built within the software that will, on every launch, report back to the company about the platform and browser that is there on the machine of the user and this data will be located in a database where this information can be mined.
In the next version of the software, this process worked out real well. The team was very enthusiastic about the capabilities of retrieving all sorts of information from the user's machine and starts building more of this. Lisa also got swept along in this tide and was an enthusiastic collaborator about trying to determine what more information can be mined. And all of this was done with the best of intentions. So, there was a need to know which printers the users were having, since this would allow the determine which printers were most often used (the belief was that more customers would be using low end printers given the profile of customers), and this would allow some more optimization of the printing capabilities.
Piracy was a big problem, so there was a thought about tracking the serial number on the user's machine, and also getting the machine name and any other such information that would allow identification of the user. Somebody had a bright idea about trying to determine whether the customers were also using a rival software, which was not easy to track, but they managed to do it by searching the registry for the other rival software. By this time Lisa was getting a bit nervous, all her instincts were reporting problems.
And then the whole thing blew up. A user who had a personal firewall and a bit of paranoia found that the software was causing a number of hits on the firewall, and decided to investigate further. As he investigated and reported, other users also started getting curious and doing investigations, and also reporting more such issues, and also raising questions on the user forums. The legal department of the company stepped in and wanted to know what all information was being tracked, and why (WHY in capital terms) they were not consulted. It turned out that there are privacy implications, and every legal team has a list of items that they find fine to track, and other info on the user's machine which absolutely cannot be tracked. The tracking of a rival installation was way beyond what was permissible, and caused a lot of problems when it was discovered.
The company did a mea culpa (not fully though), and also promised that they will ensure that at the time of installation, there will be a note that will let the users know about what all information is being tracked, and also provide the users with an option about tracking such information (this language was all couched in very positive words, but it was far more than what the team was doing). So, when you start trying to get info from the user through your software, make sure that what you are doing is above board. 

No comments:

Facebook activity