So, we all agree that data, just data by itself, is just a way to spend money in storage. Even basic analytics are an insufficient approach, since you miss the great opportunity of actioning the data. What are the benefits? Gaining efficiencies in your business operations, improving your customers’ experience and saving costs. I can’t think of a better proposition.
Let’s put an example. Imagine that you take some medical/genetic test. After that, a laboratory and you as well would have access to your health parameters, such as blood pressure, cholesterol, gluten tolerance, markers, genetic treats, etc. This is just the first step: collecting data. The second one would be when your doctor interprets the numbers and reads them to you in plain English. You have high or low probability of develop celiac disease, your pressure looks about right, but your cholesterol not that much (maybe too much cheese?). But finally, and this is what you want, your doctor should also understand your environment conditions, namely, are you sedentary, do you eat veggies, do you smoke, how many hours of sleep a day, etc., and should recommend some healthy habits and, if needed, prescribe some treatments.
And by the way, don’t think here only about curing a disease, but also about prevention and aiming for better quality of life, or even to improve your physical performance if, for instance, you aspire to be a competitive amateur athlete.
OK, enough of Grey’s anatomy.
Back in our video world, just like individuals, each playback context is unique. The content reproduced, the device used, the conditions and the reliability of the network and the business drivers of the publishers provide literally several thousands of combinations. When publishing a video ideally, you should take what I will call the “multiexpert approach”. Let me explain. You should hire an expert, someone who understands about image quality, with good appreciation for details, able to compare two images in real time and determine which one is better and by how much. That person should also understand about networks, having a map in her head about which ISP and CDN works better and how that map changes every minute. The same person should also be knowledgeable about devices, resolutions, codecs involved, image processing capacities, habits of use for that device, network connectivity and so on. Just if this is not enough, the same person, probably a PhD in several disciplines, should keep in mind what the publishing company is aiming for. Renewing subscriptions or monetize by ads? How many ads and where should they be placed? How do my audience react to different versions –with different quality– of my content?
Ultimately, our multiexpert would determine exactly what to deliver (which set of qualities, which resolution, which encoding parameters), how to deliver it (which is the most efficient codec and bitrate for the device and connectivity) and where to deliver it through. Again, thousands of combinations and only one is the best, the optimal. Oh, by the way, our expert must do this in real time, not for each one of the videos buy for each one of the playbacks (video + context) your content generates, probably thousands if not millions a month. And your expert must work 24 hours a day, never getting tired and always taking the right decision.
Like we said at the beginning, if you were able to achieve this, you could save significant costs in preparing, storing and delivering your content while at the same time your audience will be more pleased with the quality and therefore would be more engaged with it
Yes, I know, the multiexpert approach, doesn’t sound realistic. However, you can rely on Artificial Intelligence.