I had the great pleasure of hearing GoodData founder Roman Stanek speak at Cloud Connect in Santa Clara this afternoon. GoodData is just his latest undertaking as Roman was the CEO and Founder of NetBeans, the leading Java development environment, that was acquired by Sun Microsystems in 1999. He was also the founder of Systinet, a SOA governance platform that is now part of HP.
He knows the technology space and it’s hard to find someone more successful at picking the right trends.
By Roman’s estimation, Big Data really arrived in people’s minds around 2000 when the question began to be asked, “Who’s coming to my website?” While that was an important point to kick off a data conversation, it took other challenges around scale, cost and flexibility to drive the attention (and hype) we hear today.
From his perspective, what changed the game was the need for companies to build their own technologies to handle growing data volume. Yahoo, Twitter, Google, Facebook, Apache and Amazon tackled the problems themselves instead of turning to Oracle, IBM and Accenture, the traditional go-to technology companies that frankly, weren’t skilled in the problem space. This set of circumstances is what led to the odd names we throw around today like Hadoop (a stuffed elephant), Pig and Hive. These weren’t marketing terms…they were very focused tools and not built for the commercial markets.
The biggest challenges
Roman is of the mind that knowledge is the biggest challenge in the Big Data game. To use the various tools that have been home-built, it takes a body of knowledge around configuration and architecture that is very difficult to find, hire and put to use. The fact remains that many Big Data projects are really lab experiments and very hard to scale. Monetizing Big Data can seem stuck behind these three big problems:
- Popular tools were designed for internal usage
- Tool usage is extremely technical
- Success stories are very use-case specific
Knowing this, what can people do? Hire a data scientist? Roman defines data scientists as “data analysts living in San Francisco.” It gets a knowing chuckle from the crowd because we all know there’s more truth to that than jest. Data scientist is said to be the hottest skill of the future but deciding what that really means is a topic of debate.
What big data really does
We’re reaching the point where we know that data is growing explosively. Roman made strong arguments that tools, even if not commercial grade, are here to help manage the problem. But in the end, how many people are really monetizing Big Data? And if they are, how are they doing it?
Roman gave the audience three key ways monetization happens through Big Data:
- Understanding your customer touch points – Segmentation, prospect identification, campaign analysis, cross sell and up sell, retention-lapse and lifetime value
- Managing risk – If you have many transactions across many channels and products, you have to use Big Data tools to understand your risk completely.
- Mining social media data – We live in a world where customers are willing to share more and more information and that information is only useful with Big Data tools sorting through social media sources
Monetizing means deciding
Getting started has some considerations that Roman brought up as key to monetizing, including:
- Decide if you are you a zoo keeper – Managing Hadoop, Mahout, Hive and Pig means having people who understand these technologies. This is very hard and probably not your option unless you’re competing in a space such as major Telco’s or the big retailers.
- Consider investing in data platforms – Roman’s GoodData hosts for their customers but there are others as well like FedEx, Facebook, LinkedIn and Marketo that provide a platform for data of all sizes and shapes and the tools to find what matters.
- Decide if you are a Big Data platform – Maybe your needs make you a platform in your own right. You can be a data as a service provider, but realize that requires an infrastructure that has to be available, supported, and governed. This is non-trivial and Big Data platforms live and die by their SLA’s. Don’t bite off what you can’t chew.
Roman’s example of a “big data economy” came near the end of his presentation and is captured below. He made great arguments that in the end, it takes a service ecosystem that can be justified by the business case at hand. “No one would build Google Maps” was his way of describing how to look at monetizing from a very practical point of view. Cloud is what enables this economy and will only increase the amount of B2B interaction.
Tomorrow and Friday will bring more stories from Cloud Connect. If you have any questions about what you’ve read, please let me know and I’ll be happy to follow up.