As everyone reads more about big data, I get more and more questions on use cases. “How should we use big data?” is the most common question. “Are there applications available for my vertical?” and “what do others in my vertical do with big data?” are a close second on the list. There are various reports out there by many authors which describe possible or real use cases. A search on the beacon of truth will no doubt get you links to most, but are they really relevant?
Then when I was reading Fortune Magazine (I do read the paper copy – guess I am old fashioned) the graph of disconnected points in my brain finally connected while reading the edition with the Facebook cover article (volume 165, number 4). Interestingly however, the Facebook articla was not the article that got my dots connected and persuaded me to writing this post.
The article that connected the dots was about the re-invention of JCPenney. More precisely about Ron Johnson, JCPenney’s new CEO and his thoughts on how to re-invent a department store business. For those who do not know who Ron Johnson is, it is the person who – together with Steve Jobs – created the Apple Store for you. That would be the physical, brick-and-mortar one, the hugely successful one (more $$$ per square foot than any other store)…
His quote in the article (following quote is of course copyright Fortune Magazine!) says it all:
“Improvement merely lets you hit the numbers. Creativity is what transforms.” Ron Johnson, CEO, JCPenney
Again, I think that quote says it all, it tells you that you will not find the transformational use case for your business in all those articles and papers on the internet. It says that you will have to dig deeper into the actual business you are running, and try to find that one thing, that one idea that will transform your business. I’m going to guess that this idea – if you are reading this blog post – is somehow related to having more data, or doing much more analytics…
With that in mind look at the following picture. It gives you an idea of the implementations in production of big data solutions (Gartner’s Mark Beyer often uses the 5-20-75 scale to explain where technology is):
What this really means is that big data is not yet in the mainstream of technology deployments. It also means that you will not be able to buy an entire solution off the shelve, that instead, you will have to be the first one to implement that crazy successful idea. That is a good thing! It means that there is so much competitive advantage to be had by investing in big data now…
I then caught a glimpse of Oracle Profit (Volume 17, Number 1), which quotes Mark Hurd, Oracle’s President:
“Technology presents the opportunity to Transform Business”
Creativity and technology, now that is something that will really transform a business. And that is why I am writing this post! We – Oracle – can give you the technology platform, we can even give you analytical building blocks (LEGO for your analytics), but you, your business people, are in charge of that big idea.
How do you start with Big Data projects?
I hope the above at least tickled your brain, because the following will want you to believe! But before we change every business, let’s get grounded in what a real start to a big data project is. For many of us, our first big data project is going to be something that needs to prove “it” can be done.
Step 1) Gather the required technology and people to get a real project done. You will need at least (access to) the following:
- Advanced analytics (statistics, data mining, graph analytics, semantics, spatial etc.)
- A technology foundation that can handle a lot of data and allows you to analyze that data (it may not be fully at scale). Make sure you understand what technologies can do what for you, and leverage the internet of ideas to jump start your knowledge on these technologies
- A data set you know, understand, that at least supports the basic idea you are pursuing in your improvement. For example, if you are going to work on social graph driven churn analysis, you will need both your customer data and relevant social data
- A small number of people to work with and drive the above – you might not have them all ready to go, so find someone who is willing to learn and excited about technology…
Step 2) Find a problem you have today (risk, fraud, CRM etc.), that actually costs you money and improve on that problem using the stuff you gathered in step 1. This is your playground, this is your chance to learn (!) and in the process improve something in your organization. It also will prove this analytics and big data stuff really works and drives business value. Your known problem is something that gives you a quantifiable ROI. I’ve shortened the time to do X, adding $Y in our coffers by spending only $Z (where Z < Y)… Oh, and make sure you do this in the shortest possible time frame or you are going to be too late to the party.
For getting things done in #2 (and #4 below), you should actually read the Fortune article mentioned above on how Facebook works, and try to understand the mentality and method applied to new products and features. Build, show and improve or rethink but never take “no” for an answer if you believe in something.
Step 3) Build a production environment and do it all on real hardware and software. Potentially re-evaluate some of the technology you had used before (this stuff better work in production). And then do it all again in step 4 when it really matters.
Step 4) Go catch the big fish! Go after the big idea; what is the new business this opens? That fantastic new component in your services. Rally the troops and go do it! Be creative, use the technology and forget the boundaries. Do make sure you find a corporate sponsor to see this through.
Once you arrive in Step 4, think big, build the first proto-type and never, ever hedge your bets. Go for the big idea, focus and don’t try to do many other things or try to spread the risk by not really doing anything new to avoid the risk of failure.
IMHO, the way to real innovation by leveraging a big data solution is to first follow the money driven by improvement. Use that pilot to learn the technologies (analytics is key!) while solving a real tangible problem, then go all in and do the cool stuff.
PS. This is my motivational speech. I genuinely believe that most of the technology hurdles are gone, we just need to harness the creativity that naturally lives in a business, apply the technology and stick to the big ideas to transform business.