This might sound strange but we actually have a process for collecting data from the Shark Tank episodes. And the number one way is by being single.
Seriously… there’s no way you can go through so many episodes and seasons and be with someone. Wanna watch an ep of something non-Tank? Yeah… no. Or this project would never get done. So… that’s number one.
Number two is to take advantage of the formula. In seasons 5-6 (which we’ve already cataloged or are cataloging) An average Shark Tank episodes is structured as follows:
If you know this structure (and if you’ve previously seen the episode) then for our purposes, you track the following:
Money amount & equity value
(Skip rest until)
Deals or no deals are offered.
I have to admit that this is made way easier by the latest AppleTV which offers a really easy scrubbing feature on the remote. On pitches 1 and 4, I can collect the names, money and offer, scrub ahead past the demo to the sales, and then skip to the next commercial break to get the deals or no deals.
On pitches 2 & 3, I’ll scrub ahead to see how the exit looks. If it looks like a deal then I’ll scrub forward to the last deal that shows in preview and if it looks like no deal I’ll scrub to the last shark on screen. Either way, 90% of the time I’ll get my data.
The best part about pitches 1 & 4 is that, even if pitches are made in screen time that is scrubbed over, it’s always recapped in the “B” section.
Through this method, I’ve been able to capture over 350 Shark Tank deals by shortening the episodes to sometimes as little as 15 minutes.
Is this perfect? No.
I want to capture as much business data as possible. I want all the business years in sales and sometimes that comes out in the middle. As does debt a lot of the time.
But the goal of Stats For Sharks isn’t to represent the entire business, it’s to give an idea of what businesses are funded and which are not.
I’d like to dive deeper into the sales and debt numbers but they might require a second pass. So, in order to launch Stats For Sharks, we’re tracking a slightly more basic version of the numbers for the time being. But, like any good stats site, I’m planning on flushing all the data out so it can be visualized.
In the interim, though, I thought it might be interesting to have a window into how the data is being captured.