The future of business intelligence
A Q&A with BI expert Srini Swaminatha
May 3, 2018 | By Lisa Dare, TEKsystems Digital Content Strategist
Srinivasan Swaminatha is a consulting director for enterprise analytics projects for TEKsystems Global Services. Specializing in transactional systems, Srini has worked on the business and technical sides of business intelligence (BI) for 16 years, including six years at Oracle. TEKsystems interviewed Srini for his perspective on BI trends, risks enterprises take when implementing BI and what exciting developments are coming.
What trends will affect business intelligence in 2017 and beyond?
We’re moving from “coffee break” BI—when you run a query or report and grab a cup of coffee while you wait for it to process—to real-time BI. As infrastructure and hardware capabilities converge, data velocity is no longer a problem. It’s now technically possible and cost-effective to receive and analyze real-time data.
What’s also different is that users expect more from business intelligence tools than ever. In the past, BI was a rearview mirror that measured past performance. That was useful—it helped measure supplier performance, align business processes and procure items at the best price, for example. But the future of BI is getting ahead of the curve, enabling users to see whether a transaction is right before they even initiate it.
How will BI improve?
One of the ways BI is improving—and becoming more accessible to smaller enterprises—is the tight integration of hardware and software. Now you don’t have to commoditize your software and use up your hardware. With companies like Oracle and Microsoft starting to offer fully embedded software and hardware packages, BI is becoming a viable solution for more businesses.
By deploying solutions on cloud, you’re not only are making it available across the globe with a click of a button, but also eliminating the overhead, which can free up your time and money for enhanced BI offerings.
Another flood of niche products that excel in charting, graphs and visualizations are picking up pace amongst user community. Users aren’t looking for simple charts and tabular views. They want the software to guide them towards the exceptions and where to look for discrepancies; basically, they want the software to do the heavy lifting of finding patterns and identifying trends.
What do you think about deploying BI in the cloud?
Many of the security threats and ancillary questions have been addressed. Now is a good time to deploy BI in the cloud, and I see many companies are moving in this direction. Cloud solutions overcame some of the major IT challenges in an in-house deployment, including providing infrastructure support as well as keeping the core software up to date with the vendor releases. By deploying solutions on cloud, you’re not only are making it available across the globe with a click of a button, but also eliminating the overhead, which can free up your time and money for enhanced BI offerings.
When implementing BI programs, what are the more surprising risks enterprises face, and how can those risks be mitigated?
One of the major risks I’ve seen for BI solutions is user adoption. If users don’t like what they see, they start dropping from the system and using alternate methodologies. That’s a problem when you’re trying to get a single view of the truth. So you need to plan for change adoption, as well as user support and training.
You also need to ensure proper user adoption by monitoring usage—check monthly or weekly to make sure key business users are leveraging your BI tools. If they’re not, have discussions to understand their constraints and concerns and figure out how to address them.
What tools can help businesses get more out of their BI investments?
Even if you have a healthy BI solution, it’s human nature to pull the data and drop it into a spreadsheet to do the analysis, especially during crunch times like month-end. So the tools I’ve seen help companies do more are the ones that aid in performance improvement—like feeding a page that loads into the front end.
Oracle, Microsoft—all these guys have robust BI solutions, but the winner is always the one that works the fastest, which means tuning the performance to increase user adoption. It just can’t take 15 or 30 seconds per click, or people won’t use the system. You can have in-memory capability (meaning data already loaded in memory) for performance optimization.
Also, you can’t focus on all the data. You need to look at trends, or outliers. The best BI systems are KPI-oriented.
Earlier this year, Gartner drastically changed the model for evaluating BI and analytics vendors for its Magic Quadrant (MQ) report, and Oracle was removed from the MQ altogether. Do you think Oracle Business Intelligence is still a good choice for enterprises?
Oracle has been a leader in the MQ for decades, so you knew it just couldn’t fall off the Quadrant altogether. We went a bit further to inquire about background of this move. It turns out that Gartner had been evaluating two kinds of BI implementations: those driven by IT, and those driven by self-service business users. When Gartner changed its model to one that was user-driven, not IT, Oracle opted out of the review altogether.
But I haven’t seen any fallout in the market because the Oracle solution is different: it’s much more powerful because it works closely with transactional systems; it links with ERP to pull data in real time; and it can work in hardware/software mode. Oracle offers large enterprises an all-around total solution, not just niche ones. And Oracle is moving toward a cloud-enabled solution that is optimized for self-service BI, too, and we find the adoption there is picking up, especially with robust tools like Oracle Data Visualization Desktop coming on the market. From a longer-term roadmap perspective, all factors make Oracle a better choice for a lot of companies.
What excites you the most about BI and big data developments?
What ties everything together in BI is accuracy and performance. So the advances in hardware and software—the ability to turn around millions of rows of data—it’s increasingly improving performance. It’s becoming real-time oriented.
And the end-user hunger makes me really excited. You can create an architecture capable of giving additional data sources, new columns for comparison, and it’s not an elaborate IT-driven project. Once they’re exposed to successful, optimized BI, they don’t want to stop.
If you build a solution the right way—no shortcuts, but with extended capabilities and efficiency built right in—it will be there for decades. That’s pretty rewarding.