Kawandeep Virdee works on product and strategy at Embedly, managing design and direction for Embedly’s offerings and building partnerships across tech and media. He founded Hacking Journalism, a series of events bringing together journalists and developers to prototype ideas for the future of news, and he previously researched media sharing in social networks at the New England Complex Systems Institute. Kawandeep took some time to share a preview of his upcoming keynote at the annual SND Workshop & Exhibition next week in San Francisco.
You studied physics and applied mathematics in college and you researched media sharing in social networks at the New England Complex Systems Institute before joining Embedly. What got you interested in using your skillset towards studying media?
There are powerful techniques from physics research that apply to understanding societal behaviors. My research at NECSI spanned survivor search, conflict, and media. The media research was based on processing huge amounts of tweets. Seeing all of the personalities, expressions, and ways of connecting around tweets brought a lot of life into the data set. At the time, data sources became more abundant with APIs available for other media streams as well as just scraping websites. Techniques like natural language processing and machine learning provided new ways to explore our own behaviors within these data sets. It felt like an open field, and people were hungry to learn more around what could be revealed using physics to model media. This was a few years ago, and by the time I had the skill set it became known more widely as data science. For me, the most interesting questions revolved around understand and acting on our own media habits. Turns out publishers, businesses, marketers, advertisers are also interested in these insights. More recently, I’ve looked at conversation structure on Reddit, and engagement around Superbowl commercials.
Tell me about your role managing product and strategy at Embedly.
I often represent the company out at hack events, conferences, and meetups- and end up having a ton of conversations around what we make. From these conversations I get a sense of what customers are thinking about, and also feedback on what we’ve been working on. This directly informs how the products are framed, but also what we should be thinking about that we may not be. It’s that network and the conversations that I’ve found most valuable. Embedly has a range of media clients including Reddit, Medium, New York Times, USA Today and Tumblr. I focus on understanding the needs of publishers better, and addressing that in the products. For the most part it is speaking to their goals and bottom lines. The thing I always go back to is the story created around the products. I tell it over and over, and adjust it based on how it is received. Focusing on this helps keep the vision unified, and the product straightforward.
How is Embedly planning ahead for the ever-changing digital landscape?
Embedly sits between media providers- like Youtube, Facebook, Twitter, and Instagram- and media companies, which range from publishers to social networks. Being right there allows us to see trends as they unfold. What new features are emerging, and how can we bring it to a wider range of publishers? Years ago everyone became familiar with URL embedding through their Facebook timeline- and we made it easier for anyone to implement that. Now we see the growth of video formats over the last few years, and more revenue moving into video. As a result, we’ve focused on features like video responsiveness, attention analytics, and recommendations. Because of Embedly’s APIs, developers can experiment and easily test out new ideas- leading to innovations like New York Times ‘Watching’, which uses Extract. With all of the focus on video now, we’re building tools to give developers more control over players, to easily prototype new features around it. Tiny UX changes can have a huge impact. If there is any question on that, just notice how muted autoplay lead to the surge of text overlays on videos, and a lot more engagement around video in timelines. With the new video tools, it will be easier to include looping covers, include dynamic content alongside the videos, and add calls to action. The best ideas we’re not even aware of yet, but we can make it easier to companies to build them.
You founded Hacking Journalism, a series of events bringing together journalists and developers to prototype ideas for the future of news. SND has our own series of events, SNDMakes, working towards a similar goal. Why do you think these prototyping experiences are important? Any insights or best practices?
I threw the first Hacking Journalism after having conversations with editors and strategy folks at media companies. Some of the problems they were thinking about could have a quick prototype hacked together, but there was no wider awareness of these opportunities. Bringing together journalists and developers at a hackathon felt like a powerful idea. I went to SNDMakes a few months after throwing the first Hacking Journalism. SNDMakes is incredible. I changed the Hacking Journalism format based on things I found worked really well at SNDMakes. Assign teams beforehand so that each team has a balanced set of skills. Remove prizes- it’s about building to share with the group, not for a prize. Be intentional about invitations- reach out to a range of organizations and companies to have a wide range of participants. Diversity of background, race, and gender makes for a better event. After organizing a few Hacking Journalism events I’ve found that the most important thing is not the projects at all, but to bring people together and connect. You can have people at ESPN, Vox, New York Times, the Washington Post, all working together on a team, learning from each other. Bridges form across sectors too- in my case around tech and publishing. It’s about the community that is passionate about media, technology, and making things. This leads to lasting connections and powerful insights that participants bring back to their companies.
Your keynote will focus on designing informed by analytics and opportunities to respond to reader behavior. Can you share a preview of what you’ll be discussing?
Publishing is at this point where pages are dynamic- they can look back on to the audience, interact, and adjust- and innovations play in this space. When it comes to audience analytics for tech companies, everyone is trying to figure out how to measure and optimize the customer lifecycle. The main shift over the last few years is that analytics alone are not enough, there needs to be insights tied to actions. These actions can be tips of things to do, but the more effective approach builds the actions right into the product. I’ll give an example: while focusing on a publisher’s bottom line, Embedly began building out features on top of the video analytics. Analytics can be processed to drive recommendations, which boost recirculation. The recommendations are not enough though, they have to be tied into the UX. So we’re created a product that goes from measuring how videos are watched, to processing the analytics, to building recommendations, and serving the recommendations- all fitting into a simple copy and paste embed: Cards. The content is dynamics and personalized, and used last month across 30K domains with an audience of over 70M. The real values is in applying this more broadly, in figuring out how include this in designing products, and abstracting the methods to work in other contexts.