Getting into it
Marshall started off by pointing out the Forrester numbers that say that only 7% of IT and 9% of business people think that they’re getting measurable ROI from Big Data projects.
Datameer’s Groschupf provided his perspective that having a good ROI from Hadoop means taking an new approach and not simply projecting new technologies on old architectures. “Just bringing in Hadoop to have it on your resume guarantees the whole project will fail.”
Sisense’s Aziza, a self described ‘data geek’, said that, “Business users are starting to take over control.” He believes more and more ‘normal’ people are getting their hands on big data, leading to a consumerization of big data, a, “a lack of tolerance to wait for a complex solutions.”
Cloudera’s Erickson feels there are shifts going on that include a paradigm shift to Big Data. Erickson said his customers go through three phases of thought, 1) I have a problem that I can’t solve with technology like ETL and need a more scalable solution, 2) This more powerful framework allows me to store and process data incredibly flexibly with new ways of analyzing data sets, and, 3) I can build custom solutions around this, similar to what Google and Facebook did.
Marshall asked Ajit Gupta what’s he’s seeing in his world. Gupta responded that Aryaka is building a network to manage whatever’s needed in real-time. Asked for specifics, Gupta described bringing enormous data sets from wherever necessary to help, for example, mobile advertisers to move from 1 billion ads per day to numbers like 5 billion ads per day in the next five years.
Taking a step back, Aziza made the point that while Hadoop is a constant part of the conversation, while in reality big data is just the data that your organization struggles to manage. In Aziza’s mind, the definition is very elusive. Groschupf had one of the best responses to Aziza by agreeing and saying, “Big data is a big opportunity for people to make big money.” In reality, he said, the key is the time to gain insights, not the size or the particular technology.
Gupta spoke again to talk about the globalization of business and mobility. In his view, these two factors will have the biggest impact on Big Data moving forward, “Teams that are in Beijing today are working in Bogota tomorrow and they expect the same connectivity and data availability in both.”
Erickson said that what made the Big Data value proposition click for him were the stories of how long it took for companies to change data schemas to manage data they never expected to need when they created their data stores. Because Hadoop breaks apart the storage and computing side of data, it allows companies to store data up front without having to know what it will be used for.
Aziza spoke again and talked about how, “The little guys are catching up with the big guys.” He gave examples of a small companies that are part of, “…the end of sample data…” led by the cheaper and cheaper storage that allows for storage of all data. Companies don’t need to think ahead of what to store, leading to the end of the concept of how large data really is, “Users don’t care…they only care about the time to insight.”
Groschupf added to Aziza’s point by saying that the new generation of data users has a whole new mindset around making and saving money. People are using Hadoop as an enterprise data warehouse to save money and at the same time are making money by optimizing key word use, speeding download of the product to avoid losing customers to slow downloads. These were problems uncovered by Big Data analytics that weren’t in consideration a short time ago.
In Groschupf’s mind, the future is, “Storage and compute as a plug in the wall.” Rather than a focus on the the technology, he feels the future will focus on the time it takes insight, much the same as Aziza’s view.
It was a lively and engaged panel that kept the audience engaged even in the late afternoon. For a deep dive on Big Data and a workshop setting, it was the a great end to a day that covered an enormous amount of ground.