Recently I wrote about the data-driven enterprise, arguing that new data-fueled technologies, largely focused on AI and advanced analytics, will usher in the data-driven enterprise. Continuing this theme, it’s time to consider unstructured data. Surveys aplenty suggest that organizations the world over have more unstructured data in the form of electronic documents, images, video, and sound files than they do in neat(ish) packaged, structured data that generally calls a relational database home. So what? These growing mounds of unstructured data contain hugely valuable insights yet are largely ignored by BI and analytics programs within enterprises. It’s time for this to change.
Consider this: you are driving a car and carefully monitoring the dashboard instruments to ensure that you are obeying the speed limit, that the engine is performing normally, and that you’re not going to run out of fuel. However, while undertaking this careful monitoring, you entirely ignore what is happening ahead of and around you out of the windows. Seems ridiculous doesn’t it? This is at the heart of the case for bringing analytics to other forms of data, because you cannot make properly informed decisions without having access to all pertinent information.
Many enterprises are not looking at data that doesn’t fit neatly into an Excel spreadsheet or analytical tool of choice. The data may be something relatively familiar, such as the content of documents scanned and stored that show the history of a customer’s interactions with a business, or CCTV video from stores or a warehouse that could usefully provide insight to optimize the flow of people through the premises. Another common example can be found in social media. For major brands monitoring their presence online, the ability to analyze countless pictures and videos for sight or mention of their brand is invaluable.
There are good reasons why this has happened less than it ought to have in the past, when in many cases the technology to achieve it either didn’t exist or was prohibitively expensive. But this has changed. There is cheaper storage for large, unstructured data, as well as new and emerging AI technologies such as deep learning for visual recognition, and a growing range of connected sensors for physical monitoring. There are also more accessible and flexible processing capabilities in the cloud. All of these make the possibility of incorporating a much broader and inclusive range of data types into analyses more viable today than it’s ever been.
So the next time you’re looking at analytics use cases, take on the challenge of thinking outside the confines of where analytics has come from, and begin to consider what it will enable us to do with more types of data in future.
Straight Talk is a weekly briefing from the desk of the Chief Research Officer. To receive this newsletter by email, please contact us.