The failings of refugee tech

What went wrong? Not long ago, the extraordinary wave of digital projects launched in response to the European “refugee crisis” appeared to herald a new era for DSI: more dynamic, more responsive, more mainstream. But in the past few months, this appearance has faded, giving way to a more sobering picture.

The “turning point” which we were starting to feel in our first blogpost in January has veered into full-blown disillusionment. Several well-known projects in this space have packed up shop, and our research suggests that the majority of the 169 projects we know of have become inactive.

This is the first of three blogs in which we ask, honestly and unflinchingly: what happened with the “refugee tech” phenomenon? What are the serious missteps and how can we learn from them? Where do we go from here?

To very briefly recap, beginning at the beginning: Even at the time it was clear – and with hindsight even more so – that the burst of projects launched in late-2015 and early-2016 was unique in the history of digital social innovation (DSI). In Germany alone, during the peak of this “explosion phase”, there were four new projects launching each week! It wasn’t just the sheer volume of activity that was striking. There was considerable diversity in terms of who was starting these projects. The arresting images in the media of hundreds of thousands of refugees in need spurred many people into action who had no previous engagement either with DSI nor with issues around asylum or migration. Moreover, from the early days there were spontaneous impulses towards self-organisation: Slack channels and Facebook groups where this community started exchanging ideas, and several open databases (including our own) to keep track of the growing number of projects.

These are all themes I will return to in what follows, as we turn to look at the things that didn’t work so well.

But first, a note: in this blog I reference specific projects and challenges, including some of our own. In singling out projects, I intend no criticism of the people who worked so hard to build them. I have utmost respect and admiration for their vision and hard work. But though it might be uncomfortable, what is most important is that we honestly attempt understand what happened, and that as a community we learn from that. I’ll return to these points in the third article in this series, which points to some reasons to be hopeful.)

Fail #1: Building too many things

One challenge that was visible almost immediately was widespread duplication. There may have been over a hundred new projects, but the number of new ideas was much smaller. In other words, different people in different places came up with very similar ideas and started building the same thing in parallel.

When I first commented on this in an article in March 2016, I assumed that in all the frenzied activity, the projects simply hadn’t heard of one another. This was true in part, but our subsequent research and analysis showed something more complex. In most cases, projects did hear about each other early on (so the self-organising mechanisms mentioned above were working well), and the reason there wasn’t more consolidation was mainly down to pragmatic difficulties – from the technical challenge of merging two databases to personality clashes. Pooling efforts and resources turned out to be easier said than done.

One category of projects where we saw this play out was with refugee job-matching platforms. This is a case where consolidation was particularly important because it’s a model that relies on generating a critical mass of users – if there’s too much fragmentation, the danger is that no single platform reaches this threshold and they all fail. Given that eleven such platforms were launched, things didn’t look good.

However, as far as we can make out, there seems to be one – jobs4refugees.org – which has emerged as successful. A couple of the others have been rolled into jobs4refugees, but most have simply been discontinued.

Fail #2: Building things that weren’t used

Duplication was only part of the problem. A large number of the projects either struggled to get their idea to a stage where it was functionally viable, or never generated a significant user base.

There are various reasons for this, and depending on the project there will always be a degree of speculation when carrying out the post-mortem. Was too little attention paid to marketing and user experience, meaning that a potentially effective idea simply never reached the people it could have helped? Was the idea itself fundamentally misaligned with people’s needs, or was it a failure on the level of organisation or team?

Rather than looking for elusive generalisations, let’s look at a couple of examples, considering just one type of project. During peak arrivals in 2015, the situation on the ground was really chaotic. Thousands of NGOs were doing their best to receive, house and support all of the new arrivals, supported by an army of volunteers. The challenge was evolving week by week. It was a scramble. Hence, several DSI projects were trying to create an overview of what services were being offered to refugees by NGOs (i.e. mapping all social services, not just DSI). They each started to build their own database, many of them also plotting the services on a map. It appears that all of these projects have failed – albeit in different ways.

The team at metacollect soldiered on for over two years, before discontinuing in March this year. Their difficulties mostly had to do with being a team of volunteers working on the project in their free time. As such, they lacked a clear organisational structure, but moreover, they lacked the resources to execute what they set out to do.

However, the story of a similar initiative, clarat, suggests that resources weren’t the only obstacle. Despite being well funded and staffed, clarat seem to have eventually concluded that mapping a sector as large and dynamic as this one remained an elusive goal.

This fact was compounded, in clarat’s case, by the sense that their imagined use-case didn’t seem to hold up in practice. To put it more bluntly: not many people were using it. Why this was the case, and why it took so long to realise, are central the next blog in this series, in which we examine why refugees and newcomers themselves were not more actively involved in building these DSI projects.

In the meantime, let us know what you think through Twitter or email.

This article first appeared on DigitalSocial.Eu - betterplace lab is part of the migration and integration cluster at DSI4EU, a program supported by the European Union and funded under the Horizon 2020 Programme, grant agreement no 780473.

Our Podcast

The first episode of the series "Wir kriegen die Krise." (only in German)