Apps for Good

We empower teachers to prepare young people for the real world where machine learning, IoT and apps are everywhere

1. The problem

Over the past 20 years digital technology has gone from being a niche industry only impacting a small part of people's lives in rich countries to an industry that impacts nearly every aspect of people's lives anywhere in the world from shopping to health, education to love. However, the people who are shaping this industry and who are creating the digital products impacting everyone's lives today are far less diverse than the society they are now serving both in terms of gender, social and ethnic background or whether they grew up in rural areas. If we want to avoid that machines replicate and enhance our worst human biases in an opaque way, then we must educate the next generation fast to understand how machine algorithms are being constructed and what impact they have and how young people can use these powerful technologies themselves to create the type of world they want to live in.

2. Our solution

At Apps for Good we believe that a young person irrespective of their background can become a problem solver and creator. In order to achieve this vision we help educators in (primarily state) schools to change their teaching practice and to deliver Apps for Good technology content frameworks to solve real world problems. (see this image for our overall approach)

Over the past 7 years we have reached a total of 100,000 young people between 9-18 years in 3 core markets in over 1,000 schools with app development and Internet of Things (IoT) related content. Over 50% of our students are girls and a significant proportion of young people come from lower socio-economic or ethnically diverse backgrounds and many live in rural areas. We are now using our tested learning model and are extending it to the new field of AI/ machine learning. Educators access our content and support via an online delivery platform, that is then enhanced by our local delivery partners (non-profits on the ground) in our core markets that help teachers and students by offering face-to-face training and by running events. We have two communities at our heart: (1) educators in schools who deliver our course content; (2) our global community of 1,300+ expert volunteers, i.e. developers, designers, product managers working in small and big tech businesses. Those experts give advice and peer-to-peer feedback to our students on topics like technical feasibility, UX, competition, business models, in a flexible model typically via video calls.

3. How will this change the world?

In Margaret Mead’s words: “Never doubt that a small group of thoughtful, committed citizens can change the world. Indeed, it is the only thing that ever has.” When we invest in educators to train young people still in school to work together in small groups on problems they face today, they gain both the technical skills as well as the practical experience of building a solution. These teenagers then have a first real flavour of creating change and they continue to practice this competency as they grow up.  

In the education space AI/machine learning is an emerging field not many organisations are involved in yet. We completed market research in 2017 and found that until now no credible pedagogy-led player is operating in the space. We have been operating in the technology education space in schools since 2010 and what has always made us stand out is our focus on problems young people choose to work on as the main starting point, not on the specific technical implementations. In schools this problem-driven approach is still an innovative idea to many educators to date.

Technology is core to our work on two levels: (1) our course frameworks focus now on three different high-level technology platforms i.e. app development, IoT, machine learning. Specific learning tools and programming languages include e.g. AppInventor for Android, AppLab, HTML/CSS, Javascript and APIs for app development; micro:bit and Arduino/ Python for IoT; Scratch/ special extensions, Python/ Anaconda for machine learning. (2) our online platform as our core technology infrastructure with dashboards for educators, experts and account managers gives us the ability to offer different type of services to educators that we could not do without it.

Over the next 12 months (during the 2018/19 school year) our goals are: (1) to pilot the machine learning course framework successfully (2) to keep growing our number of registered schools  in our core markets as well as to grow the schools actively delivering our course frameworks (3) to deepen our reach into deprived communities by offering tailored services to schools and educators in these areas (4) to further deepen and grow our partnership work with other organisations in our space to achieve bigger scale.

Since 2012/13 we have been testing different expansion models in different countries. We realised that for our type of work the best way to scale is to work with local delivery organisations in each market. We can share all these lessons learnt and our 2017 expansion strategy review in more detail, but our new expansion strategy for the next three to five years follows basically this path in the two images below. We will grow our reach by going through these 5 stages in our current core markets and by adding new markets over time.

Recruitment of new schools depends on the local education market. For example, in the UK recruitment has been driven by word of mouth, social media, case studies of school inspectorates like HMIe in Scotland and distribution partnerships. In Portugal, recruitment of new schools is largely done by the Portuguese Ministry of Education. Our retention of schools is linked to creating new exciting course content, ongoing training and support opportunities for teachers and ongoing alignment of education policy with 21st century skills. In terms of delivery model, the chart below shows the different models used by schools across the network.

The four charts below summarise our reach over the years, the type of students involved, the impact on students and on educators. This data is generated by self-assessment surveys from students and educators. We know that this is not really academically rigorous, so last year in Portugal we also ran a control group study in partnership with the Ministry of Education, which generated positive results as summarised in this blog post: https://blog.appsforgood.org/2017/11/24/how-apps-for-good-improves-students-academic-results-lessons-from-portugal/

In the next 12 months we are expecting to sustain, but not grow our reach in the UK. In Portugal and Arkansas/ US we are expecting ongoing growth. Over the next 12 months we will continue to build partnerships both locally and globally to allow us to dramatically change our reach. Our aim is to move from reaching 30,000+ young people per year now to 2,000,000+ young people per year without growing our team nor resources at the same speed, but while still delivering a high quality service to educators.

Over the years we have built a small but effective core team that is focused on content & product development, education partner support, events, business development, industry engagement, finance & operations and working with trusted contractors on impact assessment, technical development, design. Like any non-profit we cannot pay large salaries, but have recruited and retained highly experienced staff by offering a flexible work environment (especially for working mothers), a focus on mission & impact and by growing staff who come right out of education. We have also built and regularly re-invented our way of working well with corporate partners.  

Our core revenue streams are: 1) corporates, especially technology companies 2) corporate foundations 3) philanthropic foundations 4) international social franchising income 5) income from fee-paying schools 6) government funding. 1-3 are responsible for 90% of all revenue, but year on year the percentages from each revenue stream fluctuate. We think that our revenue streams are diversified enough to manage changes in economic cycles as these different revenue streams are not in sync and we have managed ups and downs over the past 7 years.

In terms of expansion we have been very focused on driving efficiency as shown below and we do not believe that expansion should automatically mean income growth at the same pace.

Over the years we have engaged with different people at MIT who work on tools we integrated into our courses: Hal Abelson when App Inventor moved to MIT in summer 2011; Mitch Resnick and Natalie Rusk on Scratch and their recent micro:bit integration. However, we have never met anyone in person let alone engaged with them more closely. We are also excited by the recently launched Cognimates by Stefania Druga. MIT has a reach and credibility in tech education that we just don’t have as a small non-profit and joining Solve can help us build better links to MIT.

Our key barrier now is distribution, so that we can reach more young people effectively via trusted local partners. It would be amazing if Solve could help us reach that aim by looking at closer integration with MIT tools as well as helping us to build new connections.