Welcome to the third wave of business analytics
- Summary:
-
The process businesses use to dive into their data has swung from tight control to a near free-for-all. There is a better way, writes Oracle's Barb Darrow
The data analytics pendulum has swung widely over the past 40 years. In the 1980s and ‘90s the software used to crunch data was under the strict, central control of corporate IT organizations. But over time that model morphed into a user-driven free-for-all that some liken to an analytics version of “Excel Hell.”
There are advantages and disadvantages to each approach. IT lockdown guarantees that everyone works from the same, ostensibly “good” data. The flip side is bureaucracy and slow-as-molasses response times.
The free-wheeling analytics democracy, on the other hand, lets people generate data analytics reports fast, but it spawns multiple, conflicting “versions of the truth.”
We are now entering a third wave of business analytics that melds the advantages of the IT model with the convenience and immediacy of user self-service in a way that satisfies both constituencies. It offers both central control to ensure compliance and accuracy while letting users run the reports they need when they need them. Faster analytics—based on the right data—can unlock new business insights and thus new business opportunities faster as well.
Enter emerging tech
Emerging technologies—in particular, artificial intelligence and natural language processing—are key foundational technologies here.
AI, especially the subset known as machine learning, enables the software to track user interactions, automating repetitive tasks and freeing users to focus on the job at hand. The more relevant data you feed an AI engine, the more it learns how to use past patterns, based on seasonality and other factors, to recommend future actions. Making data analytics more predictive is a holy grail for businesses.
In this way, the evolution of business analytics is similar to that of the iPhone. Early versions of the iconic smartphone provided a facile interface, with subsequent generations becoming much more interactive, suggesting things to you rather than just responding to inputs.
Natural language processing makes it easier for employees who are not data scientists, or even technical, to query their data and generate reports from that person-to-machine interaction.
Most CEOs don’t know or care which corporate data sources to tap to answer a question. But they certainly know how to ask: “How did sales of Widget A in the current quarter compare to last quarter and the year-ago period? How about the previous eight quarters? Which geographies saw the weakest and strongest sales each quarter?”
A system that lets business managers interact with algorithms using natural speech brings the power of data analytics to mere mortals, perhaps pushing it past the 30% adoption rate cited by Gartner recently.
This means that tech neophytes with valuable business experience can now query their data, which can lead to valuable output. PC historians might liken this change to the early days of computer spreadsheets, which made it so much easier for people to run what-if financial analyses on the fly rather than have to spend hours reworking the numbers with paper and pencil.
Currently, a user running a modern “best of breed” data analytics product has to set up the query by dragging objects around the screen to generate charts and graphs. That approach is fine if the person knows that particular system well, but not if he or she is a newbie. It would be much easier to simply “ask” the system what a given icon or column means and go from there.
Wanted - flexible and governed analytics
Groups within companies often use a patchwork of such one-off analytics tools. For example, one tool might paint great visualizations from its own data. But what happens when a central finance or marketing team needs to sync that work with data that resides in a central IT-sanctioned repository?
Clearly, user freedom is great—to a point. But analytics anarchy breeds its own problems, in this case data sets that are not synchronized company-wide. Conversely, the risk of tight central control is that it maintains an up-to-date set of cleaned data at the expense of locking it away from the users who need it.
Oracle sees a way to meld the best of both worlds, in what it calls ‘freedom within a framework. Such a framework brings the ease of use of consumer-friendly tools into a corporate environment which ensures that access to core data is governed.
As enterprise software moves from being something that was useful but painful to use, to a more consumer-grade experience that is smarter, more predictive, and more useful, third wave analytical systems need not only to track past performance, but also offer suggestions about future actions.
But for that to happen we need both a flexible user experience along with guidelines to protect the business’s lifeblood: its data.
Businesses need centralized AI systems that come with machine learning algorithms already embedded, and apply that analytic system across all enterprise data sets. In this way, users can benefit from the power of analytics without having to guess where the primary source data is located, or having to write queries on their own. Oracle has produced an interactive “Business Analytics Assessment” to enable organizations to carry out their own peer comparison.