George Johnson
Expert Writer
June 24, 2024
It's known that Umberto Eco had hundreds of unread books in his library to understand the extent of his ignorance. Today, we can have twice as many on our smartphones and listen to their summaries in a matter of hours.
Continuing our series of expert talks with people we admire, this time we met with Denis Latysh, Product Manager for the AI RnD direction at EdTech company Headway.
Headway provides summaries of over a thousand non-fiction books on psychology, self-development, business, and more, all in clear and simple English and Spanish. It personalizes content to match your interests and uses gamification to keep things fun.
We talked about machine learning, personalizing user experiences, and the important metrics that help the Headway app succeed. This app has become one of the most popular in its niche, so we were excited to explore how Denis's team developed and grew this product.
Headway's approach may offer the right direction and innovative ideas for anyone looking to create or enhance their own app. Their strategies in personalization, user engagement, and continuous improvement can provide a solid foundation for app success.
Denis Latysh is a product manager at the EdTech startup Headway. He jumped into the IT world six years ago as a developer. Switching from coding with headphones on all day to product management felt like a big leap. Picture going from writing lines of code to figuring out what users want and testing out new ideas! Fortunately, his time as a team lead and product owner made the transition much smoother, helping him bridge the gap effortlessly.
Headway is a Ukrainian EdTech startup that got its start in 2019. They already have four products under their belt, focusing on educational microlearning apps. Their star player is the Headway app, which offers summaries of non-fiction bestsellers in both text and audio formats, complete with personalized collections, challenges, and gamification.
The niche Headway operates in is pretty unique because it lets users soak up quality knowledge quickly. Each user gets a highly personalized experience, making learning more effective and fun.
The idea for a summary app came from the team’s realization that education is evolving, and trusted, audience-approved sources are becoming crucial. So, they took on the challenge of curating great material and designing a top-notch app.
At Headway, every decision developed and tested is verified through user feedback. Even though personalization isn't a brand-new concept, iteration tests and research show that it significantly boosts product metrics, especially the Feature Adoption Rate.
So, what's the feature adoption rate? It's the percentage of users who start using a new feature after it's introduced. This metric is super helpful because it shows how well new features attract and engage users.
When users first sign up, Headway asks them to select topics they’re interested in. Based on these preferences, the app curates a personalized list of book summaries. This makes users more likely to engage with the content from the start because it aligns with their personal interests and goals.
Headway sends customized notifications to users about new summaries and features that match their past usage patterns and stated preferences. For example, if a user frequently reads summaries on leadership, they might receive alerts when new leadership books are summarized or when a related feature, like a mini-course on leadership skills, is introduced.
Users can set personal learning goals in the app, such as "read one book summary per week." Headway helps track these goals and suggests features to help achieve them, like creating a daily reminder to read or listen to a summary at a specific time. This integration of goal setting with app usage promotes the adoption of these supportive features, enhancing overall user engagement.
Headway offers reading challenges that encourage users to engage with new or underused features. For example, a challenge might involve using an audio summary feature where users need to listen to a certain number of audio summaries within a month to earn a badge or unlock a reward. This can motivate users to try out features they haven’t used before, boosting adoption rates.
Incorporating game-like elements such as points, leaderboards, and levels can make the learning process more exciting and rewarding. Users could earn points for every new feature they try or for consistency in using a feature daily. Leaderboards can foster a sense of competition among friends or with other users globally, pushing them to explore more features.
When the Headway team noticed that a lot of users had specific questions during their learning process, they decided to add an ML-driven assistant. This assistant listens to queries, gets the main point, and recommends relevant books from the library. For example, it can suggest a collection of bestsellers on time management. This chat assistant is also a key part of the app’s personalization.
A Machine Learning engineer is a professional who specializes in designing and implementing machine learning systems. They develop algorithms that enable computers to learn from and make decisions based on data, and they often work on improving the performance, efficiency, and scalability of these systems. Their role bridges the gap between data science and software engineering.
Interestingly, an ML engineer didn’t join the Headway team right away. The company kicked things off with a user-centric approach, focusing on what to do before figuring out how to do it. In fact, the recommendation system based on ML algorithms was up and running in two Headway products before an ML engineer even joined the company. Developers and analysts, knowing how crucial personalization is, researched the available tools and figured out how to integrate them on their own.
Headway's ML engineer joined the team in the fifth month of the AI RnD team’s existence when they had already validated many small proof-of-concepts. The team realized what they were and weren’t doing and saw the need for someone who truly understood the intricacies of ML algorithms. They knew having this expertise on the team would determine the app’s success.
Now, the ML team is cross-functional. It includes a UX squad (product researcher, UX writer, and UX designer) and a tech squad (ML engineer, ML ops, mobile engineer, backend engineer, and engineering manager).
From the start, the Headway app team decided to focus on creating custom covers and professional voice-overs for book summaries. They made this their main priority. Although these processes aren’t automated with AI yet, the team is exploring ways to use AI to make designs and voice-overs even more appealing to users.
The content team keeps a close eye on finish rates and feedback, responding quickly to improve content. They also experiment with the structure of summaries, text formats, and different approaches to book design, often using AI tools for testing.
To spot trends and demand for certain books, Headway studies segmented audiences by country, but the app mainly offers book summaries in English and Spanish. As a Ukrainian app, Headway also has a collection of book summaries on Ukrainian culture and history and is working on creating a Ukrainian localization to make the entire library available to Ukrainian-speaking audiences.
The best part is, if you’re in Ukraine, you can enjoy the premium subscription for free on both iOS and Android.
The Headway app team relies on user tests, moderated tests, and interviews, valuing communication with the audience as absolutely essential. For example, they might interview active users to understand how they describe Headway or how they would recommend the app to friends. This helps Denis and the team figure out what users pay the most attention to and what causes confusion or difficulties.
Using the Jobs to Be Done methodology, they gain insight into what tasks are being accomplished, what is working well, and what needs improvement.
User research usually focuses on the most active users—the team looks at the number of interactions users have with content over two weeks. A memorable story from early on is about one of the first respondents who, when asked how she realized the app was useful, said, "I turned on the first summary based on the app’s recommendation and realized it was exactly what I needed at that moment."
While analytics provide a lot of information, communication with the audience helps understand the reasons and consequences of user behavior in the app. Denis even suggests that technical specialists be present at interviews to gain deeper insights.
Headway stands out as a stellar example of how to build and grow a successful app. With a strong focus on personalization, they’ve created an engaging, user-centric experience that keeps people coming back for more.
By using machine learning to tailor content and recommend features, Headway has boosted user engagement and feature adoption rates significantly. Their dedication to high-quality content, combined with constant user feedback, ensures they keep improving and stay relevant in a rapidly changing market.
Communication with users through various research methods, such as interviews and moderated tests, helps the team understand what users really need and want. This ongoing dialogue allows them to refine the app continuously, addressing pain points and enhancing the overall user experience.
Moreover, Headway's commitment to creating custom covers and professional voice-overs for book summaries adds a unique touch that sets them apart. They're also exploring ways to integrate AI for even more engaging designs and voice-overs in the future. Isn’t that what every app developer dreams of?
And hey, if they can get AI to do the voice-overs, maybe it can also start making coffee!
George Johnson
|
April 30, 2024
Explore how AI and LLMs enhance digital products with insights from industry expert Eugene Plokhoi, Head of Product at Readdle, covering implementation strategies and model selection