I recently found that in addition to being cool for the above reasons, Stumble has a fantastic third use... cheap fast Market Research. I know, if you have ever done market research before I'm sure you're shocked that I just used cheap and fast in the same sentence <grin>
I came upon this discovery by accident.
A client had just launched a new website and they wanted us to drive traffic to it. SEO for the long term and PPC, Social Media, online PR, in the short term.
Keep in mind, this was (is) a really snazzy site. The client (who wishes to remain anonymous), invested a huge chunk of dough into the site and it looked WAY COOL. The whole team was REALLY excited to be a part of such a cool project.
We put a lot of time into planning for the site launch and did our usual stuff, reaching out to online influencers, setting up paid search campaigns, etc. Except our tactics (which are usually pretty successful) didn't pan out this time. We drove a moderate amount of traffic to the site but it didn't take off like other efforts had. This was a bit of a puzzle.
Stumble helped us to figure it out. We stumbled the site and got enough Stumbleupon traffic to make the problem abundantly evident. This really cool, hugely aesthetic site was getting a number of thumbs down!!! 
Looking at the site analytics for the stumble traffic we saw that these visitors were spending a lot of time on the home page (an average of 2 min 23 seconds) and had a very low bounce rate - 26%. We couldn't discount the experience of this traffic as being low quality, these visitors were clearly engaged and interested. So why weren't they converting? (in this case a conversion was going to a specific page and filling out a brief form)
Based on the Stumbleupon metrics, we hypothesized that although the site looked really cool, it wasn't very user friendly. Keep in mind that Stumbleupon traffic comes to the site not knowing beforehand, what it's all about. They arrive at the site knowing that it is related to their selected interests, and that others found the site intriguing. Period. No other cues such as where they came from, text links, keywords searched, etc.
Clearly in this case, the stumble traffic was confused. There were too few cues on the homepage as to what the site was about. To make matters worse, there was no navigation on the homepage just an enter here button. The button was below the fold so most people didn't see it. The homepage was really interactive so they had fun playing with it initially, but in the end they got frustrated and left.
Based on our Stumbleupon hypothesis, some pretty direct Stumbleupon comments, we recommended a number of changes designed to improve the usability of the site. Grudgingly(they were afraid it would take away from the cool factor), the client moved the Enter Here button, they added an informative title, and made it a lot easier to go in and look around.
Immediately after the changes were implemented the site metrics went up (considerably).
These were all learning's that we could have achieved through a focus group even before the site was launched. But in reality, where budget is limited and the project is usually behind schedule, who has the time or the money to do so?
Stumbleupon provided the critical feedback that the client needed. We just had to listen.