Analytics 2015: 'BI is dead: Predictive is the new black'
Let’s look at the opposite end of the spectrum to one of the hottest and fastest growing markets: the Internet of Things (IoT). Expect a ton of buzz and innovation in IoT analytics in 2015.
2015 will be the year that predictive analytics obtains significant market gains in everything from the industrial Internet to consumer devices. Here’s what Simon Arkell, CEO, Predixion Software, expects predictive analytics to look like in 2015:
Predictive analytics will be a key differentiator for modern apps: Embedded predictive analytics—whether embedded in existing applications and workflows, on devices, in memory or in real-time analytics systems—will become the key differentiator for modern apps. Consumers will expect their software to anticipate their needs, driving requirements for predictive capabilities in all apps. Spatially aware apps will become more intelligent and more common – the “who is near me” functionality of apps like Waze and Tinder will become more intelligent and prevalent in other use cases. App creators looking to embed will be faced with the key question: build, buy or partner? Leading innovators will integrate existing technology to get to market the fastest.
BI is dead: Predictive is the new black: Looking back in time to answer ‘what happened?’ is now, well…history. In order to truly benefit from data and infrastructure investments, analytics teams want to understand what could happen. Or better yet, how do I stop something bad (or expensive) from happening and how do I repeat something good (or profitable) so it happens more? The answer is predictive analytics! Some established business intelligence (BI) developers will add predictive analytics to their existing offerings, but many analytics teams will do this on their own by predictively-enabling their existing BI infrastructure and dashboards. This trend will drive the need for a broader understanding of predictive concepts and greater flexibility and usability of existing predictive analytics platforms.
New use cases for predictive analytics fuel change in key industries: Predictive analytics in marketing is nearly mainstream, with uses such as looking at customer behavior to predict preferences, next best offers and customer churn. Next year’s surprising new use cases will go beyond customer data and into internal operations:
-- For starters, let’s look at an industry that desperately needs to change to survive: healthcare. As the industry painfully shifts from a volume to value-based model, its best hope is to adopt advanced analytics, which is tough for an industry that has historically been slow to adopt new business technologies (some doctors still grumble about electronic medical records!). Clinical and administrative use cases that improve outcomes and reduce costs and penalties will generate big business in 2015. Examples include predicting patients at risk of: readmitting, having a longer than average length of stay, or even contracting a hospital-acquired infection. The opportunities in healthcare are huge, the big question is whether these historically slow-moving entities can move quickly enough to survive.
-- Next, let’s look at the opposite end of the spectrum to one of the hottest and fastest growing markets: the Internet of Things (IoT). Devices are everywhere and they are generating a ton of data, but what (besides marketing more services to you, of course!) will the manufacturers do with all this data? For starters, they should be able to predict failures. Millions of dollars are lost every year in costly downtime. If I can predict when a device is about to experience a failure and walk the end user through the steps to resolve it ahead of time, I’ve not only avoided costly downtime, I’ve created a highly differentiated user experience! Expect a ton of buzz and innovation in IoT analytics in 2015.
Speaking of devices… let’s talk about one that is just plain fun: intelligent sporting equipment. The combination of smart phone apps that track movement, steps and other details with devices that are small enough to fit in a ball, helmet, racket or bat, means sporting equipment is about to experience a major makeover. No more sitting in the locker room reviewing video tapes with your coach, when all that data can be streamed, analyzed on your phone and presented back to you. Yes, it’s possible and some pro teams are already using the technology on the practice field, but we expect to see these devices become standard equipment on the playing field in 2015.
The Shortage of Data Scientists is not over…but we’re over it: The demand for data scientists will continue to grow beyond supply, but seriously, it is time to focus on solutions not the problem? For starters, universities around the world are now offering Data Science Masters and the enrollment into these programmes as well as Mathematics and Statistics have skyrocketed. In a few years, we will have thousands of newly trained data scientists to fill the need. But not to worry, in the meantime, there are two trends that will keep data science projects alive and well:
-- Outsourcing. Ok, this seems painfully obvious, but you’d be surprised by how many organisations are hunting for unicorns when they should be turning to analytics vendors for help! Yes, it’s that simple. The good ones (hint: we’re one of them) have data scientists and other experts on staff to help you get up and running, analyze your data, find the gems and operationalize it. So, stop chasing unicorns already.
-- The citizen data scientist. Gartner gets the credit for this new term. Much like the citizen developer, citizen data scientists are springing up from within the business analyst community. These are data savvy individuals who are deeply interested in machine learning and stretching themselves to learn deeper data concepts. New tools that are simplifying the entire predictive process, are making this trend possible. Look for the democratization of data science to come as the community of citizen data scientists forms and begins to educate themselves.
Data Science tools become a commodity: First, the reigning king was SAS. Second, was the disruptive open source R. Now, Python will likely take over as the most popular data science language. But here’s the thing: it’s not about the tool. It’s whether or not what you create with the tool delivers value to the organization. As the wonder and mystery of data science begins to wear off…so will the hype (and high prices!) around the data science tools. The tools vendors left standing will be the ones who properly equipped their customers to solve major business problems leveraging predictive analytics.