Akshay Kothari, Co-Founder, LinkedIn Pulse
Starting Up with Design Thinking: The
Story of LinkedIn’s Pulse

We met Akshay Kothari, one of the two co-founders of Pulse, a business reader app that was famously displayed by Steve Jobs at the Apple Worldwide Developers Conference in 2010 and was later acquired by LinkedIn for 90 Million Dollars. Akshay and his business partner Ankit Gupta started developing Pulse in 2010 when they were still students at Stanford.

In 2009 Akshay and Ankit joined Launchpad, a Stanford design course for future entrepreneurs. The course applies design thinking as a way of gaining empathy for users through observations and rapid testing cycles with early prototypes. During that time they really began to understand the value of design thinking as a way to explore new business opportunities:

“Ankit and I decided to build something at the end of our studies that would allow us to establish a startup rather than being compelled to take a corporate job. Launchpad, the Stanford design course, was actually designed in the sense that participants would need to launch an actual product ‘into the wild’ in order to pass.

It was a fun, last course to participate in before we took off. It was especially enjoyable because we had already learned about design thinking in previous courses and had also taken a number of technical courses, which were complimentary. To me design thinking seemed to be a relatively simple framework that helped us to think about building products. We did develop products before, but we didn’t think about things like gaining empathy for our users or developing prototypes at that time. That was missing.”

LinkedIn Pulse’s founders Ankit (left) and Akshay (right)

Later during the course, Akshay and Ankit started to iterate their idea based on short test cycles. They met at cafés in Palo Alto where they also encountered their future users. It was often the little insights that shaped their idea further:

“The decision to focus on a mobile news app in 2010 came from the observation that news had not really undergone much innovation up to that time. News did not adapt to mobile. Because the solutions of all the mobile publishers were so slow, the news was really hard to follow. It was almost painful for us and for our peers to read news on mobile phones . We sensed there was something there, for us. We started observing people reading news in cafés in Palo Alto.

We basically stayed all day in a café for user tests. And yes, we were also designing for ourselves. That’s how we started. We then realized that other users felt the same way. They were also dissatisfied with how news was read on mobile phones at that time.

The iPad then came out and people started to read news from a variety of sources – but there without the content being consolidated. There were newsreaders, of course. Newsreaders back then required you to put in RSS feeds and were tedious to set up. You could not just switch between New York Times and Wall Street Journal content. Most readers looked like an inbox because they were so text-heavy. We observed users of RSS readers being overwhelmed by content. They just marked all of these thousand news as read in their settings to have this inbox zero moment. We wanted to remove a lot of that. The idea was to make it really simple for users. Pulse would never show you how many things you missed. It’s all about what’s happening right now. Pulse was also one of the early apps to display news visually. We focused on visuals so that users could discover a lot of content really quickly.”

The team discovered the power of prototyping during their iteration cycles. These prototypes changed significantly over time as the idea became more concrete.

“During class, while we were in the café everyday, we did not show our mockup prototypes right away. We first talked with random visitors about their needs and how they were reading news these days. We kept asking questions. People often say what they want. But what they actually need is something different. If we had kept doing what they asked us to do, we would have become another Google reader. Instead, we first observed our test participants as they read news on their iPads. This was a good starting point. Afterwards we would show them paper prototypes. The idea was to show multiple prototypes at the same time. People would always have a preference for one or the other prototype. People definitely gravitated more toward visual news during our test sessions. We also tried a lot of visualizations as well as different sizes and different ways to display sources. The one that is in the app right now is something we stumbled upon during our tests.

Because we wanted to test interactions and flows, at some point during our testing phase we switched from paper to keynote prototypes. Keynote helps to communicate details better than paper. We got more and more in-depth feedback. A lot of small details became important, for example the aspect of manipulation. A test subject decides that she wants to follow New York Times news and she then immediately sees visuals of NY Times news appear. This impressed our testers.

LinkedIn Pulse Early Prototype

The resolution of our prototypes increases over time as part of the process. Paper helped us to quickly test our main ideas; then keynote was the next level, and then coding was the next level. With a functional prototype at hand, another seemingly minor aspect became important. By playing around with the prototype, people asked whether there was an offline mode. This was something we hadn’t thought of before. We asked them: ‘Why do you need an offline mode when you have 3G?’ And it turned out that a lot of people read news in trains or in buses with bad connections. Thus we realized that an offline mode is actually pretty important for Pulse in the future.

There was also feedback that we did not implement because it was irrelevant to most of our users. For example, if somebody said: ‘I want to import files from other readers’, that might be a specific need and no longer the target of the app. We were therefore simultaneously gathering our users’ data to see how many of them actually demanded a certain feature. This happened from day one. You could clearly see that a lot of people never wanted to integrate news in their Google reader for example. The customers were happy to have only six features instead of one hundred. We were interested in finding more of these people for our product, instead of adding more features for the small number of power users.”

LinkedIn Pulse

As the product becomes more concrete and business scales up, the development process changes again. As new challenges arise it becomes more difficult to keep the principles alive that started the business. Askhay reflects:

“At some point we had the app and we aggregated test results of a thousand users. We had a growing team and it became harder to find time for qualitative research, e.g. going into cafés. The difficulty of performing qualitative research continued to increase. Nowadays we still invite users into our labs, but I still miss the café days. We were not giving any people any money for testing our solutions and we would get a lot of honest feedback. It was important when we were really small and we wanted to find out whether there is a market fit. At some point, however, it becomes harder to continue to do this raw testing in the wild. I guess the best insights came out of raw testing. But as you become bigger the product itself becomes much bigger as well. The question is then: ‘what do we actually test nowadays?’

We still work with principles, but it’s less structured than it was in d.school. Back then it was like: ‘now you have two days to go out and gather empathy. Now you have two days to prototype a solution. It’s not like that anymore when you run a company. I’m not sure whether it makes sense to send everyone out to do user interviews. In this respect I wouldn’t even know where to start. There are huge investments that have to be made.”

The Interviewers: Akshay Kothari was interviewed on March 9th, 2015 by Holger Rhinow, Jan Schmiedgen, and Andrea Scheer .