Why Smart Data Discovery Still Requires Smarts
Imagine you’re driving late at night on the highway. A sign saying “Last gas for 100 miles” appears in your headlights.
You glance down at your car’s digital range estimator. “150 miles till empty.”
What’s next? Do you stop to fill up? Confidently drive on?
As long as cars have been on the roads, drivers have been facing these kinds of decisions. What you have – that a driver from, say, 50 years ago would not have had – is such an accurately calculated range estimate. Your car’s advanced computer makes it possible.
But your car’s computer isn’t autonomous. At least not yet. Whether you stop is up to you.
Similar scenarios happen every day in many modern businesses, where leaders ask familiar questions like “Where do we invest for maximum ROI?” “Which product segments are underperforming?” or “Which customer group can we tap for more profit?” to name just a few.
To help answer these questions, decision makers can now rely on tools using Smart Data Discovery to unlock deeper-than-ever insights into their business data.
In this article, I’ll take a closer look at the tremendous potential of Smart Data Discovery and suggest why it’s not just the technology you should expect to be smart – its users need to be as well!
Let’s start with some background on Smart Data Discovery itself.
Smart Data Discovery Tools Automate the “Heavy Lifting” of Business Intelligence
Smart Data Discovery, also known as Augmented Analytics (more on the terminology later), involves using Artificial Intelligence and Machine Learning to automate what I consider Business Intelligence’s “heavy lifting,” i.e. the processes involved in finding data, preparing it for analysis, and extracting insights from it.
If handled manually, these aspects of Business Intelligence can be tricky for four key reasons:
-
Before you do anything else, you’ll need to define explicitly the business questions you’re trying to answer. You’re likely to overlook (or fail to consider at all) data relationships you don’t specifically seek.
-
You’ll have to write an algorithm or build a model to manipulate your data to answer your explicitly defined questions.
-
Working through the two previous steps takes time. A lot of time.
-
Throughout the entire process, you’ll contend with – consciously or unconsciously – your own biases and psychological world views, as would any human.
Smart Data Discovery tools, on the other hand, ease these sore points for users by:
-
Automatically identifying data relationships like correlations and clusters, presenting them visually or in natural-language narratives like “60% of buyers are aged 35 or over,” and making associated predictions. In doing so, your tools will be likely to uncover data relationships that would have been impractical or improbable for you to discover on your own. You do not necessarily need to have a specific business question in mind before interacting with your data and getting answers.
-
Making insights accessible via visualizations, interfaces for natural-language queries like “what percentage of buyers are aged 35 or over?” and lists of similar pre-generated natural-language queries. Voice recognition-enabled tools even allow you to ask your question out loud. In all cases, you can skip the time-consuming algorithm-writing or model-building.
-
Handling both the previous steps quickly and automatically with no significant time investment required from users.
-
Freeing data analytics from the effects of human biases and psychology.
In this way, the solutions do not just answer questions before you’ve asked them but even suggest questions you might not have thought to ask in the first place. In many ways, therefore, it may seem like Smart Data Discovery tools can do all your work for you.
Not so fast.
Smart Data Discovery Supports Your Business Decisions, But You Still Have to Make Them
In a previous article, I suggested that when all data can be beautifully visualized, the real trick will be choosing the right data to visualize in support of your goal.
Similarly, when all your business data can be “smartly discovered” and spliced together to produce countless well-polished insights, the challenge will be judging which of them truly support the decisions that matter. This extra layer of complexity – deciding what data to use to support your decisions – means that users’ critical thinking skills will be just as important as ever.
And while the powerful capabilities of Smart Data Discovery tools may make it easy to think otherwise, you will still be on the hook for your decisions. It’s up to you to have a basic understanding of how your tools work, where they support you best, and where they don’t. For example, these tools can tip you off about a number of correlations in your data that you would have otherwise missed. But will they be relevant for growing your business? It’s up to you.
In conclusion, let’s return to terminology.
Gartner, the leading research firm, recently began referring to Smart Data Discovery as Augmented Analytics. For the time being, both terms are essentially interchangeable. The name change is a good example of how even some of the industry’s best analysts are racing to keep up with its rapid pace of innovation!
What I’d like to emphasize is that both terms – Smart Data Discovery and Augmented Analytics – imply that they are one part of your overall Business Intelligence toolbox. You need a basic analytics playbook before you can augment it with artificial intelligence and machine learning. Likewise, smartly discovered data can be very helpful – but it will only make a practical difference for your business if you act on it. Use your tools wisely and stay smart.
