We’ve reached Peak Hackathon and this is what we need to do about it

Estimated reading time:

Over recent years, the term ‘hackathon‘ has entered mainstream parlance. There are many nuances in just what a hackathon is – very eloquently articulated by Jack Skinner. What I’d like to unpack today however is the growing number of hackathons in the Australian technical and entrepreneurial scene – and whether we’ve reached a point where there are now so many that they’ve become ineffective. This saturation point is a condition I’ll term ‘peak hackathon‘.

What hackathons are there?

Over the past two months alone around the city of Melbourne, Australia, there are a plethora of hackathon events;

  • GovHack – a national open data hackathon where participants leverage open data from federal, state and local government authorities to build new tools for citizens (free to attend)
  • Girl Geek Academies #SheHacks – a hackathon for women only where a number of female mentors are present, aimed at providing connections and a space for women to test new business ideas ($AUD 100 to attend)
  • #Moonhack by Code Club Australia – aimed at children aged 9-11, this hackathon is a world record attempt to get as many children as possible hacking at once (free to participate).
  • Unihack – run by Monash University’s student IT society, this hackathon is aimed at university students only and has fairly open-ended goals – with a working product being the overall goal (free to attend).
  • Random Hacks of Kindness Melbourne (#RHoK) – positioned as a social hackathon, RHOK focuses on developments that provide social outcomes (free to attend).

Why are we reaching ‘peak hackathon’?

The rise in the volume of hackathons is completely understandable. While the empirical evidence is thin – largely because hackathons are a very recent phenomena, the case studies that have emerged are generally positive – hackathons are great ways to generate ideation, to facilitate social connections and to grow innovative products and services (see references below).

Hackathons are also great was to build social capital and technical communities – for instance around a particular language, a geography or a product.

So, what’s the problem?

As hackathons gain additional traction and recognition as hotbeds of innovation, the sheer volume of hackathons being run is now becoming the problem itself.

Why? It comes down to dollars and people.

Hackathons have a number of expenses. Firstly, you need a venue to house the hackathon. Sometimes this will be donated in-kind or at a discount, but sometimes not. Then, you have to feed your hackers – well, you don’t have to, but it’s considered de rigueur to do so. Often, the venue will have a contractual obligation in place to use a particular catering company, so even if the hackathon is able to obtain the venue for a low price point, the provision of catering is often much more expensive. Next, you will need stationery, which for a smaller hackathon is often a neglible cost, but can run to hundreds of dollars for larger events. Factor in marketing and media coverage (such as promoted posts or tweets), prizes for hacks and suddenly the cost of running your hackathon can run to thousands of dollars.

Sponsorship, up until recently, has generally been relatively easy to obtain. Organisations want to align themselves with groups that represent innovation and creativity, and especially where the organisation receives additional benefits, such as the ability to scout for talent or first pick of the minimum viable products delivered at a hackathon. However, as the number of hackathons in the market increase, sponsorship is becoming more difficult to obtain in some cases. Alternatively, the amount of money that organisations are willing to direct to hackathons is diluted – meaning that a hackathon may need to deal with twice the number of sponsors – who are contributing reduced amounts of capital – thus placing an administrative overhead on the hackathon organisers – who are generally volunteers. Building sponsor relationships takes time and effort – that often needs to be sustained over several years.

An alternative to this is to charge attendees – such as #SheHacks charging $AUD 100 per participant. However, this choice – as financially necessary as it may be – places additional barriers to entry in place for participants. For instance, some potential participants may need childcare to attend (something that GovHack Melbourne provides for free), and others may need to give up paid work to attend. So, competing for sponsorship indirectly means more barriers to participation – something that all hackathons want to avoid.

The plethora of hackathons sprouting up also means that competition now exists not only for sponsor patronage, but for developer / creative / entrepreneurial attendees. A hackathon is a significant time commitment – often two-three days over a weekend – competing with leisure time, family time – or for the more introverted attendees – ‘alone’ time.  Hackathons are intense. They require significant investment of cognitive effort, long hours – and although fun, exciting and exhilarating – often leave participants tired or drained. There is a limit to how many of them attendees can actually do without feeling drained our burned out – again something all hackathons wish to avoid.

Lastly, but certainly not leastly, the other resource that becomes contended when we reach peak hackathon is volunteer time. Most hackathons – apart from corporate hack days – where the organisation has paid members of staff organise the hackathon – are run on volunteer time and effort. The number of volunteers we have in Australia has actually increased over the last five years, but the number of hours they are volunteering on average has significantly reduced. While it’s unknown whether this statistic extrapolates to hackathons and the technical community, it stands to reason that if there are more hackathons, requiring more volunteer effort, and that the pool of volunteer time is finite, sooner or later hackathons are going to contend for the same volunteers. This in turn leads to volunteer burn out – which reduces the overall capital of the community.

So, what can we do to collectively address the situation?

  • Dates – finding dates are hard. We have to schedule around university holidays (as students won’t attend if they’ve gone home for term break), major events (such as sporting events), and simply time of the year (if it’s 40 degrees outside, you might be at the beach). Trying to then co-ordinate around multiple other hackathons may then appear to be a bridge too hard to cross, particularly if the hackathon is national or international in scale.
  • Hackathon summit – another option is for the leaders of various hackathons to stay in regular and constant contact, and identify the areas where they should, and should not be collaborating. This might take the form of co-ordinating around which sponsors will be approached, or co-ordinating around dates, or co-ordinating around shared resources – for instance information on how to source childcare. Great collaboration will mean less competition.
  • Volunteer pipeline – the most effective volunteers are those who have significant experience and connections throughout the hackathon community. The downside of course is that if people are effective in a volunteer capacity they are often ‘rewarded’ with additional work. Collectively we can work together to identify, nurture and grow the volunteer base. Of course, this nurturing itself is an additional task.
  • Less money for prizes, more money for participation – with significant funds from hackathons going to prize money, it may make more sense to divert funds to participation activities – bursaries, child care, travel and accommodation grants – from prize money. Whether this would deter those hackers who come to hackathons purely for the money on offer is unknown – but it may serve to increase participation from under-represented cohorts.

What do you think? Are there other actions we could be taking as hackathon organisers to address peak hackathon?

 

Full disclosure: I’m the site lead for GovHack Geelong, a GovHack official event, and sit on the board of Linux Australia, an incorporated association which auspices GovHack as well as many other technical events such as Pycon-AU and linux.conf.au.

References

  • Decker, A., Eiselt, K., & Voll, K. (2015). Understanding and improving the culture of hackathons: Think global hack local. In Frontiers in Education Conference (FIE), 2015. 32614 2015. IEEE (pp. 1–8). IEEE.
  • Isaac-Menard, R. (2016). Hack the library Organizing Aldelphi University Libraries’ first hackathon. College & Research Libraries News, 77(4), 180–183.
  • Jetzek, T. (2016). ElEmEnts of a succEssful Big Data HackatHon in a smart city contExt. Geoforum Perspektiv, 14(25).
  • Leclair, P., & a Catalyst, O. D. I. (2015). Hackathons: A Jump Start for Innovation. Public Manager, 44(1), 12.
  • Lewis, B. A., Parker, J., Cheng, L. W., & Resnick, M. (2015). UX Day Design Challenge Hackathon to Apply Rapid Design Ideation to a Practical User Experience Challenge. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting (Vol. 59, pp. 304–306). SAGE Publications.
  • Rice, J. (2015). Hackathon implementation for industry and academia. UTICA COLLEGE
GovHack 2015 group photo, credit: Mo Xiao Xiang
GovHack 2015 group photo, credit: Mo Xiao Xiang

 

Wrap up of BarCampMelbourne 2013

Estimated reading time:

With huge thanks to Pomke and the team at Small World for lending us their fabulous Causeway House Boardroom venue, BarCampMelbourne 2013 got off to a great start.

Pomke on Angular JS

The talks I got to see where many and varied. First up, Pomke spoke to us about Angular JS and using node.js on the backend – moving processes to the client. This is becoming less of an issue with powerful browsers.

Steve on programmable logic controllers

Steve spoke to us about process control systems – and how they present a number of security risks. PLC – programmable logic controllers – drive inputs and outputs on machinery – essentially replacing buttons and switches and dials. They are programmed in ladder logic, and the PLC scans through the logic continuously – evaluating and acting on logical conditions. PLCs are low end devices in terms of capacity and communications – and could still run on 9600 baud RS232. SCADA systems are generally proprietary and generally only run on Microsoft Windows. There are risks here – such as the Stuxnet virus which was exploited before USB keys. These systems were born in the days before security was an issue, and this means that there are ways to interface remotely with many of these devices – as you have to be able to get into them to diagnose them and repair them if required. Another issue they found was that the equipment was only capable of running at 10mbps – which made it vulnerable to TCP broadcast storms.

Marc Cheong on teaching with engagement

Marc told us the story of how he became an accidental teacher – having started his PhD, falling into a tutoring role, he discovered the knack of engaging students. In his role, he found that students weren’t engaged – they wouldn’t learn anything. When exploring the underpinning causes he found that there is a paradigm shift involved in adult learning – it is self-directed, not spoon-fed. Students felt like a cog in a huge machine, and lecturers weren’t paying them very much individual attention. To remedy this problem he chose to ‘engage with empathy’ – learning everyone’s first name, ensuring icebreakers to reduce the feelings of isolation and building bonds between the class and the teacher.

He explained that the theory chained low motivation with low engagement to result in low marks – so the key to better marks is engagement and motivation – making the learning process fun and making people proud of their work.

Alec Clews on the ICT education crisis

Alec spoke about the challenges of ICT education in Victoria – there is a skills shortage, but the skills people are leaving the education system with are not great. Much of the proposed ICT curriculum should be in other parts of the curriculum and not in ICT education. For instance ethics and being safe online really belows in citizenship, while data interpretation and modelling really belongs in humanities. We need more of a focus on programming – and there was a strong sentitment in the room that visual programming is a copout. We also need more co-ordination between subjects – such as writing databases for humanities. We need to bring hacker skills into woodwork through 3D printing etc – using low cost accessible devices such as the Raspberry Pi. These devices will be a huge enabler for education.

computingatschool.org.au

Trystan on robot design choices

Trystan spoke about robot design choices, and what sort of need or objective your robot was serving and what sort of senses your robot should have. This allows you to make key design decisions so that you can build a robot to your desired budget. Once your robot has sensors, it needs some form of brain to blue all the pieces together. Microcontrollers are one way to make this happen – and you might have to design your own controller using a field programmable array (FPGA).

Lars Yencken on the quantified self

This was one of my favourite presentations of BarCamp, around the quantified self. Lars explained that everytime we use someone’s website to record something they are tracking what we are doing – but it is harder for us to capture this information about ourselves. The ideal situation would be that we have an agent measuring what we do, and providing useful advice such as ‘don’t drink that coffee because it will interrupt your sleep patterns’ based on the gathered data. Lars explained how the key areas he was trying to quantify were food and weight, but one of the challenges he had was balancing the need for bookkeeping with getting value out of doing it.

The key takeaway from me was that the mere act of measuring can serve to change behaviour – such as getting more exercise or eating fewer calories.

He also went into details about some of the glitches experienced in quantifying the self – such as battery life, GPS glitches and difficulties exporting data captured over long time periods.

One tool he mentioned that looks interesting is Huginn – which helps to measure changes in behaviour.

 

A huge shout out to to sponsors No ISP for helping make the day happen – their business model for an ISP co-operative is interesting indeed.