I was a student when I got my first internet email address in 1988. It was on a crusty old IBM mainframe that my university used for everything from statistics processing to record-keeping. I was filled with regret that I left behind four years of email when I graduated since back then there was no way to take my mail with me. Now granted, it was barely a drop in the bucket of the eleven gigabytes of mail that I've held onto over the past four years, but I've always wished that I still had all of my email.
It's for this reason that, when I got my first "real" email address in 1994, I was extremely wary about retaining control over my own mail. Along with some friends, we ran our own mail server, and I dutifully backed up my email and took care to never lose a single message. Due to my careful oversight, I now have over 15 years of email history saved, and yes, I often search back through it to look things up. When Gmail launched in 2004, I wanted to use it instead of running my own mailserver, but was concerned about once again having my mail out of my control. Of course, I quickly discovered that I could use the internet standard POP protocol to pull a copy of my mail off of Gmail for my own keeping, and I dutifully did that until several years later when Gmail started offering IMAP access, which has a number of advantages over POP (not coincidentally, IMAP is another internet standard protocol). IMAP allows me to effectively connect to Gmail with any standard mail client and pull all of my mail down to my local computer (granted, given the amount of mail I have, that might take a few days). Just knowing that I can take all my mail and move to Yahoo! Mail, Hotmail, or -- heaven forbid -- to my own mailserver again, gives me considerable peace of mind when it comes to trusting Gmail with my email. Unlike 20 years ago, there's lot of other software and services I can choose that give me control over my data.
And it's that ability to take my data with me that continues to give me a sense of comfort as more and more computing applications move to store their data in web services. At last count, I've got a considerable amount of valuable data stored across a multitude of internet servers, services and companies; from email to photos, from documents to spreadsheets, from instant messages to my address book. The thought of whether or not you can get your data out usually only occurs at the moment you want to leave a service, and if I hadn't been burned so long ago, I probably wouldn't think twice before entrusting my data to others. Consequently, I always ask three questions before starting to store data in a new service:
1. Can I get my data out at all?
2. How much is it going to cost to get my data out?
3. How much of my time is it going to take to get my data out?
Ideally, the answers to those questions would be yes, nothing more than I'm already paying, and as little as possible. If I decide to use a service that doesn't pass the test, I go in aware that I don't control the personal data that I put into that service.
Given how easy it is to try out new products on the web (and that users are more and more willing to switch web services), it seems to me that the best way to retain users in the long term is for a company to make their product so useful that you wouldn't ever want to leave. If a company stops innovating and relies on inertia and lock-in to retain their users, they're vulnerable to the next company (start-up, corporation, or otherwise) that works harder, innovates more, and just plain makes a better product for their users.
Locking you in by holding your data hostage is no longer an effective way to retain you as a user (lock the door and you'll eventually either find a way to slip it through the mail slot or break through the window and leave without your data). Many web services still make it difficult to leave their services; you have to pay them for exporting your data, or jump through all sorts of technical hoops -- for example, exporting your photos one by one, versus en masse. You should be able to quickly and easily take your data out of any product without a hassle.
It's with this in mind that we started Google's Data Liberation Front, a team whose sole purpose is to make it easier for users to move their data out of Google products. By not locking our users, we force ourselves to focus more on innovation as a means of retaining our users, and we'd love nothing more than to see every web service company follow suit.