Building systems, not just apps: key lessons for tech innovators

When people talk about digital parenting support, the conversation usually starts with content: which programme, which messages, which evidence. As the team who builds the technology, we start somewhere else. What happens when a partner needs to add storybooks to a parenting app? Or translate an early family maths app into Swahili? The answers depend on the architecture, and those decisions are made long before anyone asks.

Our work on building apps, chatbots and data solutions under the Global Parenting Initiative has convinced us of one thing: design for the next application, not just this one.

Separate the code from the content

The most valuable decision we made building ParentApp was to keep the code and the content apart. Our app builder has three layers. Developers maintain the underlying code, a middle layer holds the app’s logic, and authors write and edit content in spreadsheets without coding.

That separation has paid off repeatedly. When the nonprofit Early Family Math wanted to bring its playful maths stories and activities to caregivers, we didn’t start a new project. Its content went into the existing structure, much of it authored by a volunteer, and our Kenyan partner INNODEMS was soon piloting the app with families. ParentApp for Kids added a library of more than 50 storybooks, linked to parenting modules and play activities, without rewriting any underlying app architecture.

It also changes the usual order of app development. Normally, months of design come before anything reaches users. When the content is useful from the start, the polished user experience can come later, as the icing on the cake.

Build in every language and for every context 

Literacy only matters in a language a child understands. If adding a language means rebuilding an app, most languages will never be added. In our system, one translation workflow, run through Crowdin, covers both the app interface and the books inside it. A local team in Tanzania translated the whole ParentApp for Kids library into Swahili this way. Early Family Math is going through the same process for Spanish, French and Arabic.

Localisation goes beyond language. Partners use the same tools to adapt examples, images and stories to their own context, so a counting activity can be about laundry or digging in the garden, not something unfamiliar.

Design for low literacy, not around it

If an app assumes a parent can read, it excludes the families who most need help supporting their children’s reading. So we treat audio as a core feature, not an accessibility add-on. In ParentApp for Kids, voice guides show caregivers how to share a book through its pictures, with different voices for each character, questions and actions. Our chatbot, ParentText, lets parents choose text, audio, images or video, depending on what they can read and how much data they can afford.

The same approach applies to engagement. Small, playful touches such as badges for completing a module, quizzes and a daily prompt asking “do you have five minutes?” do real work. They help a programme fit into a busy day.

Build for the phones people actually have

It’s tempting to develop for the latest devices. The families we work with often use old, low-end Android phones, and sometimes one phone is shared by several households. We build for strong backward compatibility and offline use. Getting apps to run on some older handsets takes real effort, but it’s the difference between a tool that scales and one that doesn’t. We were pleased to see Kenyan parents gathering neighbours around a single smartphone to share stories. Good design should make that easy.

Make learning part of the product

Because every deployment runs on the same foundations, every deployment can collect consistent data and support experiments. That turns an app release into a chance to learn. In Malaysia, ParentText was used to compare four versions of a hybrid programme in one study. For Early Family Math, we plan to build research studies into the app as it spreads, so we learn from real use instead of waiting for a single big evaluation at the end.

Open stories on open foundations

Open content libraries such as Book Dash have produced thousands of stories that anyone can adapt. Combining such a resource repository with our open source app software has meant that local partners can rapidly deploy tailored, multilingual digital tools, rich in content without having to rebuild core technology from scratch. 

For technology innovators, the lesson is simple: the most important feature of an app is the next one it makes possible.

To see this work in practice, watch the Global Parenting Initiative webinars on learning through play and early literacy.

Authors

Tags

Public Health · Socio-Technical Innovation · IDEMS Impact Products · IDEMS R&D · Contextualised Variants · Deep Interoperability