Our solution
A personalized feed and topics
The core of our solution is a daily refreshed newsfeed that takes your interests, trusted sources and feedback into account. We knew that customers didn’t want to do any work. In one of my conversations, a customer said: “I’ll tell you what my business is about and you give me the gems…”.
That’s the reason why we came up with Topics. For example. The agriculture technology topic shows all the articles around this interest from trusted sources. Sounds simple? That’s the point. All the heavy lifting is done by algorithms and a - build from scratch - database that contains sources and articles.
Setting up the newsroom
Solving a cold start issue
We wanted users to experience their Aha moment as quickly as possible. However, without any input, we couldn’t return personalized recommendations. Therefore, user onboarding became crucial. Here’s our solution.
As soon as a user signs up we scrape their website. We use the text on their website or blog to find matching topics. This results in a list of recommended topics. Instead of running into a meaningless empty state, the user will see a feed filled with interesting articles and topics. From this point, it becomes fun for the user. They only need to follow topics, tell us what articles they like and hit the recommendation button.
Topics around niche interests
Organizing the noise
Topics is a unique solution in the world of news aggregators. It collects articles from trusted news sources, expert blogs, magazines, etc., and puts them into interest buckets. Not just generic interests like technology or health. We focused on niche topics like circular economy, accounting, labour law and many more. The datasets and algorithms to power this were built from scratch. It was an intense process of tagging thousands of articles and finding hundreds of sources. Customers could request new topics at any time.
Navigation
I'm only one click away
Navigation and information architecture became a pressing issue through the years. We added more features like RSS, Twitter and Kewyord Alerts. Customers followed more and more topics and added more sources than ever. This lead to frustrations. Topics and sources got buried behind tabs and nobody could find the thing they needed. Eventually we took a step back and came up with a navigation and information architecture. The navigation is familiar, customizable and brings the topics and sources one click away.
Searching articles
Fighting against a solution
As a product manager this is one of the solutions I always fought against the hardest. Customers would often ask if they could search for articles using a specific keyword. When diving deeper into their question a lot of other problems were revealed. They often couldn’t find a topic that matched their business or didn’t like their recommendations. Fixing these pain points revealed the reality of limited resources. It would require more engineers and datasets. Which wasn’t possible. In the end, we saw ‘search’ as low-hanging fruit. It was a short sprint and would take away some of the frustrations.
Article cards
Giving feedback & exploring further
In the end, it boils down to this. Users are scanning for the most appealing articles. During their search they look at the title, thumbnail, source and date. Once the eye stumbles on an interesting article there are many scenarios to take care of. All these challenges are represented in the card design. Making the newsfeed visually balanced, readable and responsive was a continuous process.
