Readability - Technical Case Study

What is Readability?

Readability helps children and youth improve their reading skills. It’s an app that acts as a 1-on-1 reading tutoring that listens as children read, and highlights and corrects in real-time any problem words that occur. 

Readability brings together Artificial Intelligence, Natural Language Processing, Advanced Speech Recognition, and Comprehension with Interactive Voice-based Questions & Answers (IVQA), Fun Content, and Real-Time Performance Tracking within one app to help young children become strong readers.

Whose needs does Readability respond to?

Readability’s award-winning Smart Reading & Comprehension Learning App gives the opportunity to get effective one-on-one tutoring that will improve their:

• Reading
• Vocabulary
• Comprehension.

What were the challenges regarding the structure and functionalities of the product?

Phase zero - Team onboarding and first fixes

Before Linnify joined, the Readability app was created as a cross-platform app. 

At the same time, development on particular operating systems had been started externally.

Linnify's involvement

1. Team structure

• 2 Developers
• 1 Designer

2. Main goal

Fix the core functionalities on the native apps ( eg. voice-based advance AI)

3. Methodology

•  Improve User Interface (UI)
•  Develop implementation per design flows

Want to explore the details of Readability app's details design?

Dive into Readability app’s design process: 

• Research, 
• UX & UI Design, 
• Prototype testing, and 
• everything in between.

Phase one - Partial redesign

In this phase, Core functionalities regarding reading and listening still needed improvements from a deeper level. The main goal was to find and implement a more suitable voice-based advanced AI.

The specific app use case needed certain tools to best fit the requirements, given the complexity of interaction and project goal.

Main limitation

Matching Siri and Google Assistant APIs to the product’s needs.

Existing challenges

1. Random hearing outside the in-app listening sessions for reading
2. Speakers switch between the ear and loudspeaker
3. The code was not optimized for scale

Solution proposal #1

We wanted to try a third-party provider that uses an open-source tool to build a dedicated custom automatic speech recognition system.

Why didn’t we proceed with Solution #1?

Time limitations

It could’ve been used to build around it, but that would’ve implied massive code rewriting. 

Since the code was not optimized for this, it would’ve decreased the overall efficiency, thus, solution #1 was not used. 

One size doesn't fit all. To achieve the product’s goals, always decide on solutions that are most suitable for the specific issue, at that specific time, with the specific resources and capabilities you have.

Solution proposal #2

Use a third-party provider with an advanced voice recognition service that works offline in real-time.

An offline tool created a: 

• Smooth experience 
• Better contextual understanding
• Sense of reliability both for the users and the developers

Yet, things still could’ve been done better. 

Thus, we started …

Phase two: Massive redesign

In this phase, the code was rewritten from scratch on iOS and refactored on Android. 

We designed an algorithm for a smoother reading user experience.

• A non-disruptive reading experience
• Specific highlighting of words while reading
• Reader’s vocabulary growth feature

Unfold the specifics of Readability’s overall challenges and solutions

Overall challenge

Outsourced team communication difficulties

• Timezone differences limit the possibility of having everyone on the same page at all times
• Discovery and mutual adoption of new tools
• Certain flow integrations imbalance between the mobile team and the backend team

Linnify's Solution

Onboard in-house specialists to help homogenise the processes:

1. Project Manager
2. Tech Lead
3. Quality Assurance

Onboarding specialists from the same team helped the consistency and streamlined the development. In-house teams assure continuous and easier communication, resulting in:

• A better technical overview 
• Quicker identification of blockers
• Faster solutioning of issues

How did the in-house Product Management improve the process?

Through clear communication, planning, and organizing, onboarding the PM:

• Increased accountability
• Brought clarity via requirements gathering
• Prioritize resources better
• Tracked deliverables transparently
• Reduced risks and points of failure
• Ensured market needs alignment
• Improved overall decision-making for the product s success‍

How did the Tech Lead improve the process?

Tech Lead gave better visibility and understanding of the processes, by:

• Facilitated the understanding of the flows by creating diagrams, and visual architectures
• Challenging unrealistic estimations
• Providing guidance, and helping to increase all team members' expertise in identifying problems and finding solutions to them
• Developing Proof of Concepts for complex features and new SDKs integration

Critical aspects

1. Build trust within the team

• Be transparent with experience, capabilities, and experience levels
• Work closely with the QA team to release confidently
• Present the flows and ideas in a clear and concise way to keep the Product Owner and team engaged in the process

‘The team focus should be on improving the product and achieving success metrics.’ - Razvan Todea, Tech Lead

2. Focus on building an autonomous team

3. Reduce Single Point of Failure (SPOF)

‍

How did the in-house QA improve the process?

The process of testing implied at least 3 layers of Quality Assurance 

• Feature testing
• Feature integration testing 
• Regression testing (final testing before release)
• The ease of collaboration between the QA and the Mobile team increases
• The overall quality of the mobile application
• Team confidence in their work
• Client satisfaction

Discover Readability's main features and the solutions to their challenges

Feature: Story Time

The app reads the story and highlights the words synchronizing with the audio.

Feature challenge

To read to the user, the app’s code requests a string and returns audio to be played. The audio file cannot tell exactly the words that needed to be highlighted. 

Linnify’s solution

To highlight the corresponding words that are read, the app needed to start a timer once the audio was played.

‍

Feature: Comprehension

After finishing reading a book, the comprehension feature asks questions from the book to test the level of attention dedicated while the app was reading the book.

Feature challenge

Given the shorter attention span of children, the main challenge was to create a valuable experience where the users feel rewarded and engaged.

Linnify’s solution

Redesign the main Comprehension feature flow in order to create a more interactive experience for the children by using advanced bidirectional AI.

‍

Feature: Parent Dashboard

What parents can see:

• Book reading reports
• Overall progress analytics
• Words/minute
• Reading session duration
• Accuracy, etc.

What parents can create:

• Multiple reader profiles
• Personalized goals for each reader 

Feature challenge

• Complex UI components
• Measure the actual user reading time
• Visualize diverse reading-related data in a simple way

Linnify’s solution

• Build custom graphs from scratch
• Time measures the active speaking sessions only
• Create robust UI components

Results

‘Linnify is a great technology partner and their dedication to our app is evident in all aspects of the product. We appreciate their attention to detail and creative approach to bringing our new ideas to life.’

- KR, Product Owner, Readability

Android Google Play | 4.4/5 stars | 383 reviews

iOS App Store  | 4.6/5 stars  | 1.2K Ratings

‍

Related Capabilities - Technology Frameworks

Native Software is designed to run on a particular operating system. Using Native development allows scalability, long-term reliability, and less prone to unpredictable issues. 

Readability App has been constructed separately on Android and iOS so that each operating system has its smoothest deployment long term. 

Native Technology Pro & Cons

Pros:

• Fast & Responsive
• Complex: has direct access to native APIs; supports multi-threading
• Long-term oriented: ideal for scale-up, suitable for complex applications

Cons: 

• Costs: minimum of two developers needed
• Extra effort in building consistency in flow and features on both apps to seem identical: needs attention to detail and good communication between Android and iOS developers
• Not a short-term favorite: steep learning curve