Relegated. Why dull soccer couldn’t save the world…

The best moments in soccer are the combinations of snappy passes, dribbles and shots that ‘unlock’ defenses. When teams are too scared of losing they risk less, the beauty is dulled and the goals dry up. Good soccer players and coaches are the opposite of dull, they’re agile.

Maybe if I’d remembered that we wouldn’t have had the plug pulled on one of the most interesting projects of my career…

This week I learned more about agile versus waterfall development approaches, and lean start ups. I’ve heard of both and have always thought of myself as a learner, and worked for organisations that piloted often and worked hard to find and integrate feedback. But I hadn’t heard the term Minimum Viable Product before, or properly understood what ‘Agile’ and ‘Lean’ approaches entail. So could they have helped?

In 2011 in the after glow of the World Cup in South Africa the non-profit I worked for was approached to flesh out and pilot a new idea. A respected researcher at an American University had drafted a concept for a public health intervention using soccer. Conditional cash transfers, employment projects, mentor-ships schemes and more had become popular across Africa. But because they often used community clinics as entry points they were almost always targeted at women. Men seldom visit community clinics and so were left out. What if the intervention used soccer structures and related incentives?

At that point our organisation had worked with more than 1 Million young people in over 20 countries using soccer as hook, delivery vehicle and community builder. We jumped at the chance but had never worked with men before, or tried to influence drug, crime perpetration or employ-ability directly. Being soccer players, not public health pro’s, we’d always been humble, and we tried our best to be evidence driven. So we built a pilot project, tested the league and activity models in existing leagues and got to work. Within a year we’d run our first pilot.

“Nobody trusted me in my family because I used to drink and smoke. Nobody gave me money and they complained a lot about how I was eating too much because I am always at home. Now they entrust me with house keys.” — 2014 Participant

A community football league recruited from bars, gangs and trouble makers. I won’t get into the details but we tested incentives like airtime, bonus league points and soccer gear, we got feedback daily and redesigned the intervention almost weekly. In the end we’d created ten consistently meeting teams who played in their own league and later expanded into local leagues in the sprawling million soul township of Khayelitsha outside Cape Town. Some teams even won local and regional cups. The model was simple. A practice day that included a basic life skills discussion, and a match day. Once up and running the teams saw 80% attendance and positive drug tests dropped from 80% to almost zero. The team’s were a self organizing support network for the men that were competing with local gangs. Very encouraging. On other indicators like gender attitudes, self reported perpetration of violence or risky sex we saw less change but the community response was very positive.

So far so good. The pilot produced some well received journal articles and conference posters and three years later funding arrived for a full scale randomized controlled trial. The holy grail of this kind of work. Money to run enough leagues to prove scalability and societal impact, and five years worth of programme and research money. A local university would run the research and recruit the participants, we’d manage the coaches and leagues. The waterfall planning began in earnest. Budgets crossed continents, a panel of experts re-worked the curriculum, the local university put an army of recruiters on the ground. By February this year 800 men had been recruited to participate and baseline interviews had been completed…contracts were in place, monitoring tools developed, launch events planned and coaches trained. Finally in March the league started, it would be phased in 8 communities at a time. The pilot had worked, the partners had nailed the gantt charts, save some nasty haggling on the curriculum we were money.

But no-one came. Day one had a less then 10% turn out. After some crisis door knocking week two was 12%. After week three the local university pulled the plug. Cancelled the whole thing. You can’t run a randomized controlled trial with a drop out between baseline and end line of more then 10%. We had an almost 90% drop out between baseline and day one of the intervention! What had gone wrong? We knew recruitment would be hard but the university said it had gone well. They’d signed up better numbers of members then we expected and had some buffer. Turns out what had happened is the researchers had unwittingly signed up men to do a survey and get the shopping voucher for their time. Not to come to soccer practice.

We’d also assumed that what had worked in the pilot community four years ago, would work in eight other communities. Looking back the pilot had been an evolving, agile play…it dribbled and passed it’s way to great results. When we tried to copy it as a block, plan and replicate it, iron out the kinks and messiness to satisfy the funders and the RCT model we lost it’s ability to be responsive. So could an MVP have saved us? What was the minimum viable product we could have tested?

It would have been easy to ask for a follow up meeting after each baseline and see how many people showed. Or to run small, cheap versions of the programme at or near a Saturday drinking spot or two. We may have gotten the message far sooner that this thing couldn’t be scaled, or even better that small tweaks earlier would mean we could plan around a lower attendance rate but still prove that soccer can save the world…