The Drudgery of the Grid, Part I
I'm intrigued by how hard it is to find and watch video online that you really love. We have more choices than ever, but it's a lot of work. Staring at endless grids of TV show or movie thumbnails is both a) amazingly cool versus the alternatives of ten years ago, and at the same time b) drudgery.
Regarding finding things online, last year I heard the word "discovery" a lot, and this year I'm hearing "curation". A recent example is the supposition that Beats Music-style curation is one of the big drivers that led Apple to make the acquisition. I think the concepts of curation, discovery, and editorial influence all deserve a careful review as they pertain to finding and consuming media online today. As users, we all need help in these areas . . . we need better experiences than we're getting from even the best services out there today. I'm going to label this Part I because my gut says that this will become a series of posts as I try to work through the challenges inherent in helping people find and enjoy great media.
What got me thinking about this was this post covering a presentation Neil Hunt from Netflix gave at a conference recently.
According to Hunt, this will change with internet TV. He said Netflix is now working to perfect its personalization technology to the point where users will no longer have to choose what they want to watch from a grid of shows and movies. Instead, the recommendation engine will be so finely tuned that it will show users “one or two suggestions that perfectly fit what they want to watch now.
Something about this didn't sit right with me. For a couple of days after reading it, Neil's comments kept popping into my mind at odd times.
After thinking about it a bit, I realized that I was bothered by it because a dissonance between two reactions I had to his comments. The first reaction was me reacting as a software guy, which was "Right on, Neil. The more users you get, the more data you have, and the more data you have and the more sophisticated your algorithms to analyze the data, the better and better the output gets. The more confident you are that a recommendation for a specific user is going to really deliver for them. Rock on, brother." But the second reaction was very different. That reaction was me reacting as a student of great user experiences, both from the design side and from the user side. And that reaction was that coming up with great recommendations is just not enough. Not. Even. Close. It's probably half the battle.
To explain why that is, think about a user on Hulu (could be on a computer, phone, tablet, or a full TV, doesn't matter). If you're designing the Hulu user experience, I think there are two key moments of truth that you need to design for. One is when a user comes to the home page and is presented with a big matrix of thumbnails, each of which will either play a video or take you to a show page or collection page. The second is when a user has watched a video and it comes to an end. Let's focus on the first one. As the user experience designer, your challenge is to get the user to choose something to watch, get them to press "play", and have them enjoy it. The last part is important. If you're just chasing clicks, you can ignore the "enjoy it" part. But if you want a long term relationship with that user, such that they love your service, trust you, and pay you with time, attention, and money, then they have to walk away loving the experience.
Thinking about this user--let's call her Shirley--facing the Hulu home page with all these choices. To get to a successful outcome the user experience has to accomplish two things. First, you have to offer choices that have a very high probability of being something that Shirley would like. Great recommendations. Sophisticated algorithms. PhD's. Neck beards. All that jazz, worshipping at the alter of big data. Great stuff. But second, you have to present those recommendations in a way that allows, or more broadly, influences Shirley to recognize that she's been offered great recommendations that are really good for her. They can't just be good . . . that by itself doesn't matter if Shirley doesn't grok that. This may seem trivial, but 6 years at Hulu grappling with user behavior makes me want to scream that it's not. Scream it with me: THIS IS NOT TRIVIAL.
Let's call this second concept merchandising. I don't love that term given its retail connotations, but despite those I still think it's the best term to communicate this idea of how something can be most effectively presented to a user to make sure they see the goodness--to make sure they actually press "play" and go on to realize the benefits of a great data driven recommendation.
So my whole point here is that you need both a great recommendation and great merchandising. And I think there has been a ton of focus and investment on the former, but not enough investment on the latter.
As an example, I often hear people saying something like "there's nothing good on Netflix anymore . . . I looked last night and couldn't find anything". I think that for most people, that's not correct. Netflix has an immense amount of content. I'm confident that for the average user, there are plenty of things on Netflix at any given time that they would love to watch. Really really love. And many of those things probably have been presented to them by Netflix, because their recommendations are pretty good. But when I hear someone make such a complaint, I think often what is missing is not better content or a better recommendation, but better merchandising. The right content just didn't get presented in the right way so that the light could go on for that user and the play button could get pressed. Or to put it in Don Draper terms, the user experience didn't plant the right itch such that the user had no other choice but to scratch it.
I think that better merchandising, and developing some science around what that means and what forms it can take, is a huge opportunity for video services like Netflix, Hulu, YouTube and others. And I think it's an opportunity for media more broadly . . . video, music, books, and text publishing on the web in general. And as users, we all need better, more efficient ways of finding media to consume that we love. Despite the wealth of choices available today, it's still just too much work.
That's what I want to explore in this series of posts. More to follow . . .