- If data is a natural resource, not unlike oil or coal, then why do companies let it pool and stagnate in data lakes, data warehouses and Hadoop clusters?
- Instead, it should flow from these data refineries to both infuse and fuel everyday systems of record and engagement.
Having spent the last three weeks on the road, attending user conferences from Salesforce.com, SAP and IBM, I’ve learned two important things. First, sleep is a precious commodity, often hard to come by and truly something not to be wasted. And second, enterprise data is likewise as precious a natural resource that, like coal and oil, should not to be wasted. Of course, unlike true natural resources, data is anything but scarce, bubbling up from systems of record and engagement, flowing freely from server farms and permeating the ‘Internet of Things’ (IoT).
So, why do we spend so much time as an industry focusing on the best way to stockpile and refine data within deep but often impenetrable data lakes, aging data warehouses and unkempt Hadoop clusters? I’m not suggesting that these means of storing and processing data aren’t important. On the contrary, you cannot build a ‘System of Insight’ (a phrase coined by IBM which I quite like) without understanding the data driving that system. And you cannot gain said insight without employing a means of querying that data.
There is no shortage of data visualization and discovery tools on the market capable of diving deep into any data store to retrieve pearls of wisdom. There are even some solutions capable of using that data to suggest which questions to ask. That is the case with IBM’s emerging Watson Analytics cloud service, which aptly combines machine learning, natural language processing, domain expertise and predictive algorithms to literally guide users to insightful outcomes.
However, such solutions shouldn’t be an end unto themselves. We should not view systems of insight as insular, omniscient oracles like Arthur C. Clarke’s Hal 9000 or D.C. Fontana’s M-5 (for all you Star Trek fans out there). Where such machines can do the most good is not in the realm of the macro, but instead within the very systems they aim to improve — traditional systems of record or the emerging, decentralized systems of engagement that drive discrete business objectives.
Cognitive tools like Watson Analytics should bubble up insight within line-of-business applications themselves such as sales enablement, customer service, fraud detection, enterprise planning, etc. That’s where they can do the most good, at the point of action, where real business users must make tactical, real-world decisions that directly drive revenue.
That’s where developer ecosystems come into play, as with IBM’s fledgling Watson Developer Cloud. My hope is that such efforts will succeed in encouraging developers to embed cognitive systems such as Watson Analytics, or any data discovery and visualization solution for that matter, directly within line-of-business applications. That’s where they can do the most good.