As engagement becomes a more valuable and measurable metric (and we know through many test across verticals that increasing engagement leads to increased conversion and page views) optimizing flash video movies and other rich Internet applications (RIA) like Ajax become critical. I’ll explore Ajax optimization in an upcoming post. This post was prompted by a recent all-day client brainstorm for a site with LOTS of flash video. So let’s focus on the original RIA that user experience pros love to hate, Flash.
As a background we have been optimizing flash for sometime with clients. In fact a number of clients use Offermatica not only as an optimization and testing platform for Flash but an analytic tool as well. This is because Offermatica scripts can count action inside a flash application unlike many other analytic tools. This becomes very important if you use flash as a strategy to engage user actions, the only useful reason to have Flash in the first place and the key reason that it needs to be optimized. So without further adieu:
6 Examples of Flash Video Optimization
The best video is not the one the interactive designer or the brand agency delivers. The best video is the one that gets the highest percentage of users to perform an action.
Example A: Travel site. Do you show the video with the smiling kids and family enjoying their time together or do you show fantastic clips from some of the most popular travel destinations? How about a mix of both? Certainly with such divergent themes one creative would perform best. Of course it’s possible that users hitting the site from searches for “family vacations” might be influenced by one version and searchers that had queried “Paris vacations” might prefer another. More on targeting below.
If we have the right theme we next need to find the right length based on the metrics we are trying to improve. This one is interesting in so many ways. As a lover of YouTube and self-professed optimizer, a question I’ve wondered about for a long time is the optimal length of a YouTube video to reduce drop-off and maximize video (page) views per user. Of course content quality will impact this metric but that has no bearing when the content is the same for all users (or users within a segment). It also gets mitigated the more content you include in the test as well as when the content is segmented (e.g. the optimal length of a YouTube video in comedy is likely different than film & animation).
Attention optimization is incredibly fascinating. Gord Hotchkiss is leading this thinking in search results with his eye tracking work. At OTTO Digital I believe we are at the forefront of work in attention optimization for digital video. Still the best advice on video length optimization might be from Charlie Parker, “always leave them wanting more.”
Once we have the right theme to capture attention and the right length to keep attention then we become story tellers with editing.
Example B: Apocalypse Now would have been a great movie had been edited any one of a number of ways. The script, photography, acting is all great. But even as great as it is, there is likely another cut of the film that would have appealed to a wider audience or different cuts that would have appealed more to different audiences. I would bet just about anything the film could be cut a different way that appeals more to women than the original.
I subscribe the belief that every message has an optimal way to be shown so that it appeals to the most people overall and a diferent way among particular segments. To me this is one of the most interesting and emergent areas in digital film and video. At its core it questions the role of artists and audience in the creation and consumption of art in much the same manner emergent technologies and networks threaten the the digital advertising status quo.
Bottom line: The “audience cut” of your flash video means that you have improved what you set out to do in the first place.
This is a pretty basic one. In most cases sound is still something most users do not expect. With a strategy to meet and exceed user expectations this generally means no sound. However in certain instances it is expected and can provide benefits. The only way to determine impact is to test.
Example C: Site in question has most of its daily audience coming during working hours yet they default to having sound. Probably a bad idea worth testing. Sending the cubicle alert out that you are surfing during working hours is a sure sign of quick page exit. But maybe a large percentage of users don’t have speakers or work with headphones?
Choice and Control
User control of the experience we will increase engagements. A good thing. Too many choices and the perception of difficulty will decrease engagement. A bad thing. Still, what are these choices and controls and how do they affect performance. MLB.com is a typical rich experience site with a homepage center well of videos, six in all with one default. Are six the ideal number? What about the order? The number (choice) and order (control) can be optimized based on actual user interest.
Example D: Take a control percentage of traffic; say 20% that gets a mix of presentation order and number of videos. Based on the actions and results of the control group serve the remaining 80% the optimized delivery and presentation. Better yet, adjust it in real-time. We’ve done this many times with links and products but the same practice can be applied to any content including video (one client has actually done this). I refer to it as content merchandising.
As with virtually all creative the most important optimization technique is targeting. We know that images and video have an incredibly strong capacity to elicit emotions, the core of branding and an incredibly persuasive element in conversion.
Example E: In the MLB.com example above geo-targeting the default video would be a great optimization technique. Sadly, when I hit the site today I received a video on the Red Sox and A’s and not the Yankees or Mets (geo-targeted based on my New York City ip address). Presented with choices of Yankees or Mets I would have clicked on the Yankees. Having expressed my affinity I could be retargeted by MLB.com on repeat visits by defaulting to the Yankee videos that are much more likely to engage me and get me to return to the site in the future.
If you have or will invest money, time and resources creating and delivering flash video for your site and have given it prominence in the site experience to be front and center then how can you not be optimizing it? Especially considering that it may be hurting site performance rather than helping it. Flash video does not have to be the enemy of users. If the presentation and delivery are optimized it can provide a richer, more relevant experience to users that will improve the results of your business and the value of your brand.