Case Study 2— Zach Schendel, Director of UX Research at Netflix

Read Case Study 1 here, in case you missed it https://medium.com/@punam.garu/case-study-1-zach-schendel-director-of-ux-research-at-netflix-15c6f57f181a

Hi, everyone! So recently I attended this webinar organized by InVision with Zach Schendel of Netflix Design Team and I found it very informative and interesting.

In the webinar, he discusses the evolution of Netflix’s design and how they went about in deciding the design changes. I am sharing my notes on the webinar here. I have basically organized the information here so that I do not miss out on anything good.

I will be posting 3 blogs to cover this webinar, segmented across the 3 case studies that he discusses.

Case Study 2

In case study 2, Zach discusses what is personalized imagery and why they decided to have them in the bottom 1/3rd of the screen space.

Images on the Bottom Row

Research Method — Eye Tracking Study

Netflix conducted an eye-tracking study with Netflix users, where they made participants come to a lab, and wear the eye-tracking glass with a camera, gave them a remote control and asked them to find something new to watch. Something new means not just new to the world, but new to them. Something that they have never watched or heard before.

The focus here is on Content Discovery.

Eye Tracking Results

Findings:

1. Adults are reading the content unlike kids we saw in Case Study 1
2. Most of their attention is on the first row of images below and mostly they are looking at the images
3. The browsing process is very quick, with an average of 1 sec on each image
4. The viewers will choose the path of least resistance while choosing a show to watch

People spend 70% of their total browse time on that top row, particularly looking at the images

They rejected 91% of those titles simply by looking at those images for roughly 1 sec

When the user hovers on the title, the video sometimes takes a sec or little more to start playing. But people browse these titles so fast that a lot of times even before the video starts playing, the users have already moved ahead.

Users keep browsing very quickly unless something might just catch their attention and then they might pause a little longer to watch the video play.

This research was done around 2013–2014

Action Steps

Image optimization

Images were not algorithmically optimized around this time. Everyone would see the same images no matter what is their browsing history.

How it was done?

Netflix took some titles and created 6 images for each title including the current image (baseline image) that was shown to the users.

Then they showed these images to users to understand if the baseline image was the best image, or if there was potentially a better image out there, which can increase engagement with the title.

Above, the baseline image was the top left one. Results showed that the bottom right image increased the engagement most — relatively 62% more engagement than the baseline image.

Result

With this Netflix was able to give people globally the best image out of a group of 6 images.

Ethical Standpoint

Q — Is Netflix tricking people into watching these shows?

Ans — No. Netflix is only including quality plays here. Which means if the user is viewing a show for a significant amount of time (definitely above 5 mins), only then Netflix attributes their viewing to a particular image.

Improvisation

Netflix now creates around 10–15 images of each title. And a user sees that title image, which people with a similar viewing history to his, have engaged with mostly.

So all the images are personalised according to a user’s viewing history.

Inference

1. Understanding where to focus your design efforts is very important
2. Netflix could have spent time and energy in optimizing the top section, which covers almost 2/3rds of the screen space. But then it wouldn’t have been as impactful in improving the user engagement
3. Understanding that people are putting more focus on the bottom 1/3rd images row to choose the show they want to watch was an important finding
4. The decision to optimize and personalize these images in order to help people choose the show they want to watch is a good example of focussing your design efforts in the right direction
5. This arises from user empathy, where you actually feel the users pain point in choosing the show. Understanding that they are putting in time and effort to find out a show, and helping them in that direction is important.

Balancing Design and Utility

Always understand that the design you are creating will be used by real people who will try to accomplish tasks through the designs. Accomplishing tasks is equivalent to accomplishing goals and they create a sense of fulfillment and satisfaction in the users' minds. Which is directly proportional to the amount of engagement in your platform.

For Netflix users, this task includes scanning through loads of titles, including Netflix Original series and movies which the user has never heard or seen before (never seen in movie theatres or other channels). The user has to go through all this unfamiliar content and then decide if he wants to watch this show or not.

1. So you can create a design that looks amazing but doesn’t give the user the kind of information they need to make that choice.
2. Or you can create a design which has all the information but is not placed correctly for the user to view it. The content is placed in a way where the user doesn’t pay attention to because it’s not easy for them to look at those locations as compared to other locations.

Hence conversations and knowledge sharing with the design team are important, where the research team can share notes on where people are actually seeing while they are browsing content so that the design team can find a balance between designs and utility.

Watch the full video here — https://www.youtube.com/watch?v=ld7RfeZNtXs&feature=youtu.be