What good is a prediction if its not based on data? When I look at our own internal data and the larger big data market in 2014, there are a few interesting trends that emerged that I believe are strong indicators for what will come in 2015. For an in-depth explanation of 2014 trends and my 2015 predictions, you can watch this webinar, but I will share a few of the highlights here.
Looking at our own data from last year, the first trend we noticed was a shift in titles of those investigating big data. Over the course of 2014, there was a marked shift from IT to Business, where in the second half of the year business executives far surpassed their IT counterparts in investigating big data offerings.
Second, we saw a vertical shift in interest, where traditional early adopters like ad-tech and media and entertainment companies fell off as they’re well on their way with big data, and more business user focused, use-case driven sectors like business services, retail, financial services and healthcare interest skyrocketed.
Third, we moved past a lot of the more common use cases like website behavior analytics, and saw increased interest in newer use cases like fraud & compliance.
Based on these observations, plus hundreds of conversations we’ve had with both IT and business leaders, I believe we have the following to look forward to in 2015:
Big Data will be a business initiative. Historically, what we saw was that IT tried to service business, and built the infrastructure for them, but its clear now that business users themselves want to get their hands on big data. It’s no longer that business sends IT off to find analytical results, the line of business wants to do that themselves. There is a really strong demand for the democratization of data.
Hadoop deployments will shift from being centralized in IT toward departmental deployments. We also saw that central, IT-driven deployments are shifting to department-specific deployments. Unfortunately Hadoop is not yet to the point where business users feel confident around the security requirements that certain business units have, like HR, or other areas where personal identifiable information is involved, so what those departments are requiring are separate environments, and not shared services. You would think the benefit of Hadoop is a centralized environment, but clearly there is still some question marks around multi-tenancy and security, and in 2015 I believe those questions will get even louder, and we’ll see a response from the Hadoop vendors.
The conversation will shift away from tech and toward business value. While we’ll still hear plenty of conversations happening around technology differentiators, the above-mentioned business driven demand will shift the conversation to what business benefits can be achieved with big data. Additionally, while there will be some very honest conversations around TCO, the bigger focus will be on ROI.
All of the above trends, when combined, really leads to this next prediction: Hadoop As A Service (HaaS) will become prevalent as more companies want to get big data analytics up and running faster. New lines of businesses that want to start their own big data projects quickly, will do proof of concepts in the cloud, and some will stay in the cloud. Others will certainly move to on-premise, and as a whole, increased cloud adoption will only further accelerate the overall adoption of Hadoop. Cloud will enable a lot of people to accelerate time to value, and that is really the most important thing.
Finally, I will leave you with a teaser, as this last prediction really deserves its own blog post. We believe one of the most critical things that needs to happen for Hadoop in 2015, is that Hadoop needs to convert from a special purpose platform for big data to a general purpose platform for any data. As I mentioned before, with all the business users that are coming online with Hadoop analytics, whether its with Datameer or otherwise, for all of those business users, it’s all just data to them. They don’t care if its large, small, structured, unstructured; for business users its just critical to get access to the data. That can’t truly happen until Hadoop makes the technical shift from a special purpose platform to general purpose platform.
That, briefly, is what we’re looking out for in the big data market in 2015. I’d encourage you to watch our 2015 predictions webinar to get the full picture and explanations, and I’d be interested to hear your reactions or questions in the comments below.