Innovation Indemnified: Creating a rapid learning and a faster experimentation culture at your workplace

So here's the deal. I've worked in multiple organisations to date and have heard the term 'innovation' been thrown out as the saviour for all lacklustre thinking during company off-sites and planning meetings. We live in an age where established organisations offer 'innovation time' for their employees. For e.g. Google offers it's employees, 20% of their work time to work on what they think will most benefit Google. They basically offer their employees time to think. At the rate at which mobile is taking over our lives; having thinking time is a  bigger necessity than a 'nice to have'! As Rand Fishkin's recent tweet suggests, mobile is not killing desktop, it's killing our free time. 

After working for Intuit, I really thought I'd never be able to join another organisation that can really change the game in the innovation space. Intuit not only offered official innovation time but they also had an innovation catalyst programme and also encouraged small 'pizza teams' that ran on the whole design for delight principles as well as the lean start up principles that helped employees experiment and get deeper insights to inspire for innovation.

Thought Intuit was cool? Then comes WSO2 in my career and changes the game completely. WSO2 is a lesser known brand than Google or Intuit but are becoming a leading innovator in the tech space. If you can't believe my word, then just have a chat with the folks at Gartner. 

But what I really want to talk about is not what the Gartners or the Forresters of this world think or what the hundreds of customers such as Boeing, eBay, Fidelity, HMRC etc...think about WSO2. I want to share a very personal experience of mine at WSO2 as an employee.

The Open Email Culture: Any WSO2 story wouldn't be cool if one didn't know about the email policy. WSO2 products are 100% Open Source and the culture within the organisation is very close to being 100% transparent. The emails are a great example of that. At WSO2, there are multiple email groups and most of these groups are visible to almost all employees. And it's highly encouraged to participate in any email thread that you deem you could add some value to. So more often than not, a colleague at engineering might look at one of my email threads and shred apart an idea or a fellow marketer might drop a line or two adding a bit more insight into how better we could execute or else our CEO himself would add his thoughts! (Yes it does happen!) What we are encouraged is to get as many ideas, thoughts, criticism as possible during the conceptualisation phase of an idea. This helps one to think alongside many aspects of an idea and also pits the idea against an internal set of experts that can sometimes sound like a promising VC or a bunch of investment dragons. But at the end of the day, once there is a quorum for the idea, it's up to the owner to carry it through. The beauty of this is not just getting a +1 or a -1 for an idea but actually getting a chance to learn from one another. 

Collaborate and Learn: The email groups are just part of this whole innovation process. What this does is, it creates a hub of knowledge that can be tapped into when ever needed. Collective knowledge is far more powerful than any individuals' and an open culture such as this definitely encourages that.

The #BigDataGame: If you were a patient reader so far but really wanted to know the true meaning behind the cover photo of this post; then now's the time to be very excited! This is a great example of what an open culture within an organisation can bring about. From the day I joined WSO2, I've been blessed with an awesome team who have helped churn some great ideas into reality. The #BigDataGame was one such idea. I've always been a massive fan of sports in general and have been following some of the major sporting events and the impact they've had and continue to have on consumers; especially from a brand recall perspective. As most of you may know, a 30 second advert on the big game day would cost around $5million bucks! So does that mean the world of marketing and advertising is for Goliaths of this world? Hell no! 

Whilst I was thinking about this, I got a very interesting email from  #ThinkwithGoogle; which ironically got me thinking more! The email had some pretty interesting search related stats about the Superbowl and how the search war for the SuperBowl began a month prior to the actual event and how post event marketing also helped to get lesser known brands out there.  WSO2 is known for great technical content and more importantly for it's kick ass open source platform. The Superbowl on the other hand is usually full of fluffy marketing and $5million commercials. But how could one stand out in a fluffy marketing environment? DROWN, in heaps and heaps of data.. BigData in fact!

Here's a video tribute to the big game by some of the WSO2 guys!

https://youtu.be/kc4IUUk96ww

WSO2 has a cool platform and a very interesting stack of products. One such product is the WSO2 Machine Learner. WSO2 Machine Learner takes data one step further, pairing data gathering and analytics with predictive intelligence: this helps you understand not just the present, but to predict scenarios and generate solutions for the future. Whilst not specialized for sports, Machine Learner is all about generating predictive models.

So in true WSO2 fashion, I dropped an email to the marketing email group as well as to the engineering email group seeking additional thoughts, advise on my idea, and voila, got a few immediate replies including a few party poopers :) But more importantly got a response from one of the Tech Leads and a dear colleague from the Machine Learner team who was actually interested in exploring this. But he could only offer me a bit of his time and a full time intern for a few weeks. This was more than enough to explore this.

The predictions landscape and the geeky bit (few excerpts from the WSO2 blog): After a few idea exploring sessions, we looked at the NFL analytics space that was out there. Nate Silver (god of modern big event predictions) and the founder of fivethirtyeight.com was leading the pack with some pretty cool Elo based forecasting models. But we thought ours could be better especially because of the awesome stuff that's behind our Machine Learner product. We used Random Forests - generally one of the best classes of classification algorithms out there. We paired this with stacked autoencoders, which is basically a layered neural network where the outputs of each layer are connected to the inputs of the next layer.

This worked. After comparing it with actual values, we realized the algorithm was highly accurate: 96.7% accuracy, to be exact. Unfortunately, the model only generated 1s and 0s; if both teams had the same value (1 or 0), we were simply unable to make the prediction. This went on in this vein for a while. The logistic regression algorithm, minus the encoding, was okay (96% success rate), but only gave us boolean outcomes. We mucked around with the features for a while, including trying out multi-class classification algorithms, which were, unfortunately, not as accurate as we wanted for this exercise. We’d reached the limit of what we could do with our current crop of Machine Learner algorithms.

This wasn’t surprising, because everything in Machine Learner was geared more towards business and enterprise data prediction than, sports. But rather than letting it die, we decided to build a pure Apache Spark program for the random forest regression algorithm we needed. We then ran that to try and predict the 2015 dataset.  

We did a little bit more calculation and arrived at a mathematical 76.5% accuracy rate. While not as high as some of the previous models, this one isn’t binary; wins are predicted as percentages in themselves. That’s very important, because this system of odds is how humans also think about sports. This gives us very high hopes for the model.

The Execution: Whilst the Engineering guys were working on the tech side of the project, some of us looked at ways in which the predictions could be used in a cooler way. So the self service drag and drop predictor was created. Whilst the design team was busy working on creating and hosting this project; guys from the content, legal as well as the sales teams chipped in with great suggestions. It also helped having a few major NFL enthusiasts in the sales team! Mind you, all this on that one email thread! This email thread was full of energy and it felt as if no one wanted to let that energy die; especially when we came across some of the legalities of this project.

Through this great collaboration effort, we managed to launch the #WSO2 #BigDataGame just in time for the playoffs and at the time of writing this blog post, we managed to accurately predict 7 out of the 10 play off games! And the ones that we couldn't predict, were mostly plagued with injuries! Now, this program isn’t perfect. It’s very much a hobbyist project at WSO2, and we’re still in the process of tweaking it. Right now, we’re basically predicting probabilities of success - and while we have faith in our product, there’s a whole lot of things that are impossible to account for, injuries in particular. There’s also no predicting the effects of morale on a team; that stuff is sorcery.

This has been one hell of a ride and it feels great to have been part of this super awesome team that developed the #BigDataGame. Our little project was even featured on CIO.com, PCworld and Computerworld and up-voted on Quora and Reddit and has brought a significant number of organic traffic to the landing pages.

The unwritten rule at WSO2 is that we don't carry arm chair quarterbacks. We operate like a start up when it comes to ideas. And this was one great example of how a great open culture facilitated in bringing about the best in a set of passionate people to get an idea off ground and implemented in double quick time.