Time for a retrospective on the Firehose hackathon for Coca-Cola

The critical measure is what happens when the hackathon is over. People give so much. Time. Energy. Imagination. In the case of Firehose, we had over 20 mentors, 5 judges and 50 participants giving up 2 and half full days, plus a week in between all on the promise of learning news skills, building something that customers want and meeting amazing new people.

Learn new skills. Build something with 3 years worth of Coca-Cola Amatil’s supply chain ‘big data’ (that’s every delivery, case, customer, truck, etc for the past 3 years of deliveries in NSW, VIC and TAS) using a top tier database platform for big data: SAP Hana

Create a startup that someone wants. It sounds simple, but most hackathons build easy stuff that is quick to validate in a couple of days. We wanted to build early prototypes of startups that could become huge new businesses for epic companies such as Coca-Cola.

Get to meet some amazing people. Build your network.

We created 12 new startups over the course of the event. Almost all of these used the Coca-Cola dataset on Hana. We provided training before the event occurred but, even so, the learning curve was huge. But people did it. We got there. Coca-Cola Amatil will continue to support more than half of the businesses, 3 of them have early interest from investors and one of them has followup meetings with Toll (who attended as mentors and sponsors). We learned stuff, and we built businesses that have a better chance than most of going to the next level. Here’s an idea of the startups that were created.

We learned some valuable lessons

Love. A hackathon creates startups out of an impossible set of constraints. There is limited time to think and we can’t rest on the crutch of convention. This creates the environment to share a new way of working with teams that is reinforced with an emotional pillar. It is such a pleasure to see the rush of excitement when a business idea is validated with a real customer just hours after inception. This is a place where we can make work that we love.

What’s in it for me? Before inviting anyone to a hackathon, we must ask what’s in it for them. Participants, mentors, judges. The ‘transaction’ must be fair and clearly articulated.

Customer validation makes a huge difference. Using the lean startup process and emphasising customer validation in the judging criteria creates more prototype businesses out of a hackathon. Without it, teams can gravitate to build solutions that no one wants.

Themes are critical. Establishing a theme creates the alignment between participants, sponsors and mentors to create an experience that everyone values. It also prevents participants from bringing along a pet project that alienates the rest of a team that join them.  Everyone is in the same moment, creating something new from nothing. Themes really do determine who comes along. At Appiness, which focused on charity solutions, we saw a bunch of women participants, while Firehose attracted mostly men and also attracted data-scientists and engineers who had not tried a hackathon before.

Make it human. Abstract themes are hard to work with. Introduce the human side and solutions come alive. When we introduced the themes for Firehose, participants heard from one of the truck drivers who clearly and passionately shared his problems that technology could solve. Not surprisingly, half of the startups created responded to this theme. We did a lot of this at Firehose with tours of CCAs distribution centre at Eastern Creek and opportunities to sit in at the call centre and see the trucks.

Tool up participants for what happens next. Many startups born at a hackathon fail because they get stuck in IP and ownership discussions between founders and organisers. At Firehose we were clear. All IP sits with the team that created it until someone does some other deal. This is fair as it is them that are taking the entrepreneurial risk. We also provided talks and templates for the founders to have the right discussions about equity.

Things we will do better next time

Scary NDA. We needed participants and mentors to sign an NDA around the big data set that CCA provided for the event. It was a scary document that needed a lot of trust in us to sign when people arrived at Firehose. Next time we will try to make documents like this simpler, explain them better, and give people notice.

Unlocking mentor value. Next time we will publish more information about each mentor to describe how they can best help teams and we may even give teams a way to book in that specialist help. This is to minimise time wasted with participants and mentors having long discussions trying to find a fit.

We surveyed the participating teams on the event. Survey Results.

A judges perspective from Annie Parker – Co-Founder of Muru-D, Telstra’s new accelerator

Last weekend I was asked to judge the Firehose Hackathon in Sydney.

Hackathons are a tough thing to get right, how do you ensure that the startups come up with meaningful ideas that solve real problems? How do you ensure that the startups get value out of the process, that they learn something? How do you ensure that there are tangible next steps?

For me, the Firehose Hackathon went a long way to solving those issues.

By seeding the hackathon with real world data it meant the startups could focus on fixing actual problems rather than just creating something random.

Coke didn’t just give data though, they gave access to people from the whole value chain; call centre workers, drivers, head office staff – this way the startup teams could qualify their ideas & check they were on the right track all the way through the process.

I was really impressed with the quality of the ideas presented to the judging panel & many of them were easily transferrable to other industries as well — 2 of these ideas were that good I’ve offered personal coaching time to them to help shape them going forward.

To those who think this is just a way for corporates to steal ideas from startup land I disagree. Hackathons are one of the most meaningful ways corporates can engage the startup community to share with them the big issues they have & focus the extraordinary talent on fixing them.

Could there be more done? Sure – there could be dedicated programmes of future involvement for the very best ideas from hackathons promising longer term support to help shape those ideas into real businesses, we could be sharing that data to the startup community on a regular basis….but Rome wasn’t built in a day was it?

A mentor’s perspective from Pratibah Rai

When I signed up to be mentor at Firehose, I confess I did with reticence. My background is mainly in consumer startups, what could I possibly be adding to big a data hackathon? But since Phil had extended the invitation, I put faith in his decision and said yes.

And boy am I glad I did. The first weekend reinforced for me that lean methodologies can work for all new businesses, be it consumer startups or big data ones. Finding a problem-solution fit can really only happen with customer validation, and no business wanting to ‘make it’ can avoid this crucial step. What was special in this scenario was seeing the validation happen in real time. Coca-cola had made key customers available to the hackathon participants, and most dived head in validating assumptions. It also quickly became obvious that teams that didn’t use this valuable resource suffered when taking their ideas further.

The big moment for me personally was the visit to the Coca-Cola warehouse in Eastern Creek. The sheer enormity of the scale of operations was …. I wanna say mind-blowing. During the tour, one of the things our guide, Mark Hopkins, said struck a chord with me. In an operation as immense as Coke’s, improving a tiny piece of the mammoth process leads to exponential benefit (time, money, efficiency) down the line. The process that they have from start to finish works. What is needed is small tweaks within the process to make it better.

That’s when the synergy of a hackathon and a corporate clicked for me. In my past hackathon experience, the teams that do well are the ones who have found a fault in a process and produce a solution to tweak it. And the reason their solution impresses is they have recognised the time constraint of a hackathon and focused their efforts. This is exactly what Coke was looking for from a hackathon, and I think they got it. The one difference between this and other hackathons is, from what I’ve heard, there is real potential for some of the startups pitched on the final night to go further and really ramp it up. And that’s really exciting. Because I think that startups and corporates need to start playing together here in Australia to affect the ‘big picture’. It’s a key partnership we need to make work to get the attention of government and a chunk of the Aussie economy.

Organiser’s perspective from Franki Chamaki, Co-Founder at Coca-Cola Accelerator

The biggest challenge in any startup is scale. In order to gain scale, startups either need big marketing budgets or align with a strategic partner to give them a leg up. For strategic partners to develop, “networking” is critical. With the recent Sydney Firehose event, around 50 participants had the opportunity to do just that. They didn’t just build businesses over a weekend, they built key business relationships and trust along with shared value. In addition, they accessed real tools and amazing resources not readily available to a startup to help them validate their idea. They also worked on real problem while owning all the IP that was created.

See what Franki did next as part of CCA’s commitment to open innovation. 

Organiser’s perspective from Mark Hopkins, Logistics Project Manager from Coca-Cola Amatil

Key points:
-SAP training on a new tool that the corporate world barely knows about. (20 participants from the entrpreneur community)
-tour of a working CCA warehouse with 2 hours of question and answer with staff.
-participation at our national contact centre listening on delivery support phones.
-in addition to two winners and a people’s choice another team has a meeting with Telstra Muru-D and different team has a follow up meeting with Toll. This all before we even have the next CCA pitch session where I hope that at least one more of the teams will get endorsed to progress further.

Our staff was completely energised by the event and were really pleased to get to meet so many interesting people that can develop actionable ideas quickly. The partnership between corporate and entrepreneur is an opportunity for both parties. CCA was really pleased to see how much we can benefit from partnering with startup community and believe that our dedication to the event and follow up are an example of that.

Participant’s perspective from Ian Hansel, co-founder of Auto Ordered.

What a week! Coca-Cola Amatil (CCA) decided to take a gamble and let 50 entrepreneurs, designers, developers and data scientists dive head first into their data at the firehose big data startup competition. Why did they do this? CCA wanted to see what insights a fresh set of eyes could bring to optimise their supply chain. For the hackers it was a chance to get their hands on some real world big data, learn some new skills and meet other people interested in the startup scene.

The first day of the competition was focussed on market validation, testing out ideas and data mining. Pollenizer had a real buzz for the whole day. It was a long day and most teams changed their idea considerably from what they first envisioned. Lots and lots of learnings. The competition continued throughout the week. While Auto Ordered was hacking away at the data and asking customers what they thought of the product other teams were going to Eastern Creek distribution centre, riding in trucks with the drivers and going to call centres. All to get a first hand look at the problems they were trying to solve in the supply chain.

Following weekend and it was crunch time – slide decks were written, then rewritten and re-rewritten. We improved our pitch considerably after Chris Petersilge’s pitching workshop. This demonstrated the benefits of taking people through the customer’s experience and making an emotional connection with the audience. Something us analytics folks can forget from time to time.  All the pitches were of a very high standard – hard to believe none of them existed just a week ago! Auto Ordered took out the top prize with a system that uses ordering history to predict what a customers order should be. Some of the other winners included:

– CokePass, an awesome cashless payment system

– Night Rider, a way to sell off excess truck capacity at night

– Big Data Dynamix, an interesting big data analytics and visualisation solution

In conclusion it was an amazing event. This was my first hackathon and I wasn’t sure of what to expect but it was a fantastic week of learning how to quickly turn an idea into a validated product. Real world problems solved with real world data. In addition to the cool things that were built it was great to meet so many other people with a passion for creating useful products with data. Hats off to all the organisers for making the event run so smoothly.

—

Gerald Berkovics from the runner up, Big Data Dynamix, also blogged his experience.