Is a Big Data Conference Like Going to the Dentist?
I recently visited my dentist and found the experience, well, less than satisfactory. I was subjected to 45 minutes of lecturing while strapped into the dentist chair:
“You don’t brush enough times a day. You don’t brush long enough. You don’t brush correctly. You aren’t using the right type of toothbrush. You aren’t replacing the brush head soon enough. You aren’t flossing enough. You aren’t flossing correctly. You aren’t using the right type of floss.”
Good lord! It’s any wonder that my teeth aren’t just falling out of my head!!
No one likes being lectured. It’s not constructive and after a while, you just turn it off. That’s probably similar to the feeling of going to these Big Data conferences – constantly being told what you aren’t doing right. And maybe I’m guilty of doing that as well. If I am, then I’m sorry because from an audience perspective, it sucks.
So the purpose of this blog is to highlight what organizations are doing right (and I’ll save my criticism until my next conference presentation). Here’s what I’m seeing organizations get right with respect to Big Data and Data Science:
- Organizations, and especially management teams, are getting comfortable with scale out, natively parallel, open source-based technologies. Organizations are moving away from rigid monolithic applications (e.g., ERP, CRM, SCM, HRM) to an operating model comprised of special-purpose applications. It’s like being a carpenter who only had a hammer and screwdriver their whole lives. Suddenly they have saws and tape measures and wrenches and planes and more. And if they find that they need another tool, they can even build it themselves. Unfortunately, this bounty of options has made the CIO’s life much harder. A CIO can no longer just buy a monolithic application from a single vendor and then get comfortable being held captive to that vendor’s ever-increasing maintenance pricing and ever-lengthening release cycles. CIO’s must trade the perceived security of the “one throat to choke” lie for the agility and scalability of assembling their own application environment. It may be a lot more work for the CIO, but the technology agility and business value that will be released is truly game changing (see Figure 1).
- Organizations, especially Chief Financial Officers, are getting comfortable with open source business models. I suppose that we can thank Linux and Red Hat for introducing the open source business model to the corporate world, but many developers embraced the open source phenomena many years prior when using the Free Software Foundation’s GNU development tools (e.g., gcc, g++, gdb, cvs). Now large commercial organizations such as Google, Facebook, Pivotal, Dell, and IBM as well as many universities are adding new functionality to existing open source projects, and creating new open source initiatives (under the Apache Software Foundation banner) whenever they can’t find an open source tool that sufficiently meets their needs. The pace of innovation is something that NO monolithic application will EVER be able to match! And that’s a fact, Jack!
- Many organizations are hiring a Chief Data Officer (CDO), which is a great start. Unfortunately, many of these organizations are hiring the wrong person and putting them in the wrong role (sorry, but a small lecture coming). Many organizations are hiring CDO’s that look and perform like a CIO. Hey, we’ve already got a CIO. Instead, what we really need is a “Chief Data Monetization Officer” whose role is to drive and derive business value out of the organization’s growing bounty of data. In fact, in some leading organizations, the CIO actually reports to the CDO. Now that’s really embracing the power of data! See the blog “Chief Data Officer: The True Dean of Big Data?” for more details on the upgraded CDO role.
- Understanding the difference and complementary natures of Business Intelligence (BI) and Data Science. Business Intelligence cannot be “extended” to include data science, and data science does not replace Business Intelligence. They are separate disciplines and must be treated as distinct and different entities. BI provides operational, management and compliance reporting on what has happened, which is the foundation for any data-centric organization. Data Science leverages the growing wealth of internal and external data to try to predict what is likely to happen, and prescribe what actions to take (see Figure 2).
See the blog “Dynamic Duo of Analytic Power: Business Intelligence Analyst PLUS Data Scientist” for more on the differences between Business Intelligence and Data Science.
- BI Analysts and data scientists alike are embracing the power of a schema-less data management environment. It’s really hard for us data warehouse people (who were taught to think schema first) to transition to a world where we just load the data into Hadoop regardless of how the data is structured. Semi-structured, unstructured, video, text, photos, images, audio, log files…you just don’t care. Understanding and embracing the power of “schema on query” versus “schema on load” truly is game changing, especially if you want to do some serious data science predictive and prescriptive analytics. See the blog “Hadoop Data Modeling Lessons – by Vin Diesel” for more about the ramifications of data modeling within a big data environment.
- If everyone has access to the same open source software, then where is the point of differentiation? Well, the differentiation is NOT in having the open source software, but is instead in how the open source software is being used to drive and derive business value. Organizations are realizing that this conversation must start with the business. The “Democratization of Big Data” means that just having deeper pockets is no longer enough; that even small organizations with a tight business focus can have impressive success (see the blog “Heart of the Data Science Revolution… Mankato, MN?”).
- Finally, I’m seeing more and more organizations discussing how to “monetize their data.” However, we need to shift the monetization conversation away from the “value in exchange” discussion to the “value in use” application. Understanding the true economic value of data is a game changer, and will change how organizations optimize their technology and business investments (see Figure 3).
Check out my recent University of San Francisco research paper on the “Economic Value of Data”.
Ultimately, your big data success is dependent upon your ability to answer these 5 key questions:
- How effective is your organization at leveraging data and analytics to power your business models?
- Do you understand your organization’s key business initiatives and how they would benefit from big data?
- Do you have business stakeholder active participation in setting your use case roadmap?
- Do you understand the economic value of your data and how that affects your technology and business investments?
- Do you understand how to create a platform that exploits the economic value of your data?
Now I hope we’re really talking, and not just lecturing at each other!