My key learnings from building Apps for Good

When I leave my job with Apps for Good at the end of August, it will have been exactly 10 years after I completed the initial feasibility study. While there are 168,000 charities in England and Wales according to Charity Commission data, Apps for Good is part of a much smaller group of only 7% of charities with annual revenues higher than £500,000. This in turn has allowed us to impact a total of over 170,000 young people since starting out. 

However, when a charity founder and long-standing CEO leaves there is typically just silence – on the successes and failures that must inevitably have happened. I think that the reason for this is that until today most people working in the nonprofit sector fear that funders are unforgiving when things don’t go according to plan. While this is certainly true for some charitable funders, what I have experienced over the past decade is that most funders we work with do understand that when you try something truly innovative and new only a small part will initially work and that you must iterate and pivot to really achieve the impact and scale you aimed for initially. Just like in any other sector. I deeply believe that the charity sector as a whole can only keep getting better when we share our key learnings.

So what are mine? 

1. It is critical to have all skills and capabilities for a nonprofit in the core team early on. As a technology education nonprofit our biggest gap in the core team was that we did not have someone with digital product experience who had created a successful user experience or online service before. Hence when we started hosting our learning content online in 2010 and started building our first online educator, expert and student user experience and platform in 2011, we made all the mistakes we could possibly make in that area. It took us years to build up this expertise inside the team and it cost us a lot of time and money over the years. On a positive note, we had a very strong fundraising capability in the core team right from the start and soon after we built a strong finance function at board and team level. As a result, our fundraising and financial management have been very effective during the last decade, allowing us to manage funding cycles, to launch new income streams, to control costs and to ensure a stable cash flow. 
2. We were right to pivot our main channel away from informal learning centers to schools. When you read the initial 2009 feasibility study and compare it to today’s Apps for Good operating model there is at least one big difference: schools are not mentioned – at all. In 2009 I was convinced that schools were bureaucratic and impossible to enter as a new education nonprofit. Hence I focused all my efforts on understanding the channel that had worked best for CDI in Brazil: youth and informal community-based learning centers. After the financial crisis hit and the new coalition government had to cut spending in 2010, one of the key areas impacted was funding for communities and local government on which local youth and community centers depended. As a result, many youth and community centers shut down or reduced opening hours. My initial channel for reach did not work. On the other hand, after the initial showcase of our first students in March 2010 and some press coverage, we got approached by schools and launched our first after school club at Central Foundation Girls School in the East End of London in September 2010. Working with professional educators and schools ensured that we had a channel that really worked to reach young people. When Debbie Forster joined the Apps for Good team in spring 2011 we then also had someone as part of the core team who understood our new channel extremely well. 
3. As a free education programme we found it very hard to strike the right balance between control and openness/ flexibility. When you run an open education programme that gets educators and other education nonprofits excited, you quickly face the challenge of how to retain some control over brand, values/ core principles, delivery and impact. This is not because we are control freaks, but because as a charity we are accountable to our funders who rightfully want to know who their money is reaching and whether it has an impact. In the early years (2010-2013) we solved this problem in the UK by creating barriers to entry in order to assess whether schools interested in Apps for Good were actually serious about delivering it and we stayed closely in touch with them via school visits, face-to-face training and regular phone and email contact hence building strong relationships with them. This approach did ensure that we had a fairly good knowledge of what was happening on the ground, but it equally meant that we turned away many interested educators and schools at the time because we could not support everyone with our small team. When we did lower the barriers to entry to accelerate our UK growth in 2013-2015 and to take significant market share we were not fully prepared for the additional data tracking needed, the change in relationship with educators and schools and the necessary changes to our support model that would result. When we implemented the change at the time it was not clear to me that this would really impact every aspect of our operations and required us to reassess every aspect from a cost benefit analysis. We learnt that under certain circumstances face-to-face support and training are still justified in an open model as personal interactions create high impact – despite being more expensive. 
4. Technology learning content has a short shelf life and must be constantly refreshed to retain its value. After the initial creation of our learning content in 2010 and small improvements after the pilots, we did not refresh and update our app development course until we were forced to do so in summer 2013. While in other education domains it might be totally ok to update resources only every few years, in technology education new learning tools regularly emerge, existing links and tools become outdated and ways of working in the technology industry change that must be reflected in your content. Since summer 2014 we refresh our existing courses once a year. However, after running app development courses for five years we also realised that apps were no longer as new and exciting to educators and teachers as they were in 2009 and that newer technologies had emerged that had not yet covered in our courses. Hence we had to diversify our course offerings. With hindsight we were too late launching our new course on Internet of Things (IoT) only in 2016, but with the launch of the Machine Learning (ML) course in 2018 we were finally back at the leading-edge of bringing new technology into the classroom. The ML/AI education space is fast emerging and we can play a key role so that young people solve problems that matter to them – just this time with ML prototypes. Ultra IR, machine learning powered ultrasounds detecting problems in foeti, winner of this year’s Emerging Tech category at the Awards, are a great example for this. 
5. Europe lacks a cluster of philanthropic funders focussed on investing in strategic international expansion of nonprofits. Thanks to the demand for our programme internationally, we started working on growing our footprint outside the UK from autumn 2012 onwards. While we have a well-oiled fundraising process in the UK, where we know that we have a 25-30% success rate when submitting a qualified proposal, internationally our funding success can be best called erratic. It’s not that we did not manage to raise funds for international expansion at all, but the amounts we raised were between £10,000-£100,000 and were mostly linked to specific geographies that funders were interested in. Despite trying several times in different ways over the years, I did not manage to build a solid pipeline of funders that invest strategically (i.e. not country specific) in international expansion of nonprofits. I know that there are exceptions where some private European foundations do these deals behind closed doors and where high-net worth individuals sign such a cheque, but I was not so lucky. In the US, as far as I can see the landscape is a bit different, and there are more funders that do these types of deals and are more transparent about it. It’s also worth noting that there are European-based nonprofits that have expanded internationally, but when you look for example at Amnesty International it’s clear that most funds come from individual donations i.e. are based on a B2C (business to consumer) nonprofit business model, whereas Apps for Good and many other education nonprofits rely on foundations and corporate donations i.e. operate a B2B (business to business) nonprofit business model. 
6. Our move to a flexible and remote working culture has been crucial in growing and sustaining a strong team. When we started Apps for Good our working culture was close to what most commercial tech start-ups operated at the time: a small team crammed in a shared office working very long hours on a low salary. With the big difference being that as a charity we could only entice people with the prospect of generating significant positive impact on young people and educators, not a big cash out. After a few years into Apps for Good we realised that beyond pay and mission there was a third asset in our then changing working culture we could build on: flexible and remote working. I think it definitely helped that I moved to Berlin in 2014 and with the exception of a monthly visit to London became a fully remote worker. Using this asset proactively we were able to recruit and retain highly skilled staff who needed flexibility in terms of location (remote) or working hours (part-time). Our flexible and remote working culture has by now benefited everyone in the team at different times. Just a few examples: your heating is broken and needs to get fixed, so you work from home; your family is based abroad and you can work a few days from there; you have child-care responsibilities and only come to the office once a week. We keep refining how we use our remote infrastructure, i.e. Slack, GDrive and Hangouts/ Skype. Today social banter happens online rather than purely across the desk. When you read about start-up working culture today in 2019 we are again part of a much bigger trend in supporting flexible and remote work. 

Without a doubt I have more learnings from the past decade, but I hope that these six resonate with some readers. It has been a great journey for me running a small charity with a start-up mindset. I am keen to see the next chapter of Apps for Good unfold and will give it all the support I can from board level in the future. And by joining Junior Achievement, the world’s largest entrepreneurship education nonprofit, as Director of Technology and Innovation in their European office I will stay involved in the nonprofit sector for the foreseeable future. 

Iris Lapinski, Founder, Apps for Good