This week, I installed the latest beta from Firefox. A bug in the process somehow resulted in Flash no longer working within my browser. I tried to fix it,to no avail and, thus, I surfed the Internet without Flash for the last few days. (Yes, I know. I could have just used Safari, but I am making a point here. :) )
For the most part, my experience wasn't that different. All sites loaded fine, thought every now and then a Flash-based header wouldn't come up. However, for a while now, most design gurus have advised against using Flash in front page user interfaces, so you don't encounter it much.
That said, what was most obviously missing from the Web for me during the last few days was video. Without Flash, there just isn't any video on the web. Sure enough, as I read through blogs, visited news sites, and combed my RSS feeds, there were big obvious holes where video had been embedded but didn't apparently exist without Flash running. It caused me to reflect back on the dramatic impact Flash has had on the realm of Internet video.
During the early days at Speedera, customers used to come to us to host their streaming media events on the web. We ran a large network of servers with lots of available bandwidth and could handle the peak traffic loads they expected. So, big TV networks, movie studios, sports leagues, and even tech and consumer product companies would come to us to ask us to dedicate our network capabilities to enabling massive amounts of user traffic to their streaming media events.
The way our payment for our services worked back then was that a customer would establish some traffic estimations and then commit to buying that much of our service. If they had user traffic above that amount they would need to pay extra. And, in the beauty that only results in inefficient markets, if they used less than that committed amount, we would keep the overcommitted money nevertheless.
The most intriguing artifact of this system is that we got deep insights into the variety of models our customers would use to predict their web traffic for the streaming media in question. For the most part, the estimates would take some sort of known number of off-line audience (e.g. how many people watched a TV show on a given night) and then discount that for how many of those folks would then visit the website and watch some streaming video. Invariably, this form of analysis would bring us customers who were absolutely convinced their streaming event would see 300,000 users.
It was 1999-2001. These were the heady days of the Net. Everyone knew its promise, and each company with whom we dealt had a VP of Digital Strategy who felt personally responsible for demonstrating this explicit potential to their stuffy gray-haired colleagues. These guys were tragically confident that they would put up record-breaking traffic numbers during the actual event. Much to the chagrin of some of our salesguys, we used to try to talk customers out of agreeing to unreasonably large deals. However, the customers often insisted on ponying up gargantuan sums, as if there were some pride on the line. No one wanted to be chatted down from the edge of an event that would surely land them a cover story in The Industry Standard.
Yet, the cold hard truth was that, though the technology existed to stream video on the Net, no one was watching it. Sure, Mark Cuban had traded Broadcast.com off to Yahoo, however, the mainstream was still not watching online videos. We used to joke that when a customer came in asking for a traffic allocation to cover X users, we should subtract three to four zeros from the estimate and provision for that infinitely smaller number. With only a couple notable exceptions (e.g., Steve Jobs keynote addresses, Victoria's Secret fashion show, both delivered by our competitor Akamai), Net video was just not being experienced by the masses.
Why? It was too hard. Remember those days? Multiple players spanning from Real to Windows Media to Quicktime. Did you have the right codecs installed? Were your clients assigned to the right filetypes? The right version of the player? Did your firewall allow a file to be streamed? Buffering? Still buffering? Timed out?! Argh!
Video on the web simply did not work. Content guys were doing deals to make features available, but it just didn't matter. Clicking on a link triggered an often unpredictable race of client software to open and attempt to play it. It was like rolling dice and very often you came up a loser. No clip to watch, just a broken experience. As a result, only the savviest users could be counted upon to watch online video.
This all changed with Flash. Kevin Lynch, now Chief Software Architect of Macromedia, once told a room full of us at Google that Flash is actually more ubiquitous than Windows. A Google engineer recently reminded me that, beyond that, it very aggressively auto-updates and prompts for installation if you visit sites with Flash on them. With Flash everywhere, it makes for such a wonderfully universal platform for rich media accessible by the masses.
The Macromedia folks seized this opportunity and built a wonderfully easy system to stream rich media. Flash video content could be seamlessly embedded in a page and just one click would allow it to play right there in context. Wow! Video finally worked! Flash made it so easy. No more downloading ugly incompatible clients. No more worrying about wacky DRM regimes and trying to keep versions current.
Flash showed that if you make it easy enough the users will follow. So, while there are many reasons to explain the success of Youtube, and I do I admire those guys, in the midst of all of this deal frenzy, I hope someone pauses to write a little thank you note to the Flash developers for making video on the Net finally work.