The beginning of business intelligence started in the 1980s. In-house IT departments started doing batch processing of data at the end of the day or week for manufacturing resource planning.
This soon expanded to become enterprise resource planning (ERP). But as ERP systems wrangled with more functions – including accounting and finance – more and more data was generated. IT departments became overwhelmed with requests for more and more updated information.
“We have clients who want to do advanced analytics but they have never done any historical reporting . . . You simply cannot go from that to advanced analytics no matter what your vendors tell you”
Fast forward to the 21st century where business folks expect up-to-the-minute information, and it becomes clear something had to change.
“The typical IT guy would take anything from six to eight weeks to add a data source,” explains Bhavish Sood of IT research firm Gartner. “Business users will say, ‘I don’t have the patience for that’. By the time you add that, the business would have already changed.’”
He adds: “Business guys are now using more and more analytics products.”
But business users are complaining. “If you look at a typical business user, he’ll have three main gripes with BI. First, there is a lack of self-service: ‘Every time I need to write a report, I have to go through the IT guys.’ Second, it takes too long to fetch data.”
“Third, there is no visualization. The usual pie charts and bar charts are doable, but if I want to do complex visual presentations, I have no ability to do that.”
On the first two points, the need for access to updated data and ever more complex presentation has led to a shift in BI that gives the business user increased flexibility and closer to the ideal of self-service. The advent of cloud computing has made it possible for IT departments to move away from the command-and-control model of information access to a more fluid one.
Warning: You Need Data
New technology has also made it possible to use data to do predictive work, but Sood warns against having unrealistic expectations.
“We have clients who want to do advanced analytics but they have never done any historical reporting,” Sood says. “They have never done any kind of BI. You simply cannot go from that to advanced analytics no matter what your vendors tell you.”
“You need a couple of years’ worth of data to do model testing and calculations. If you’ve not had any data for the last five or six years, it is pretty much going to be impossible for you to do any meaningful prediction.”
And that data should be reasonably clean. “If you have data quality issues, the report you produce will not be good,” says Sood.
“What I’ve been advising clients before they start a BI initiative: make sure you’ve done profiling of your data, and that you have a model for data enrichment and fixing data issues. Otherwise, you’ll be live on the BI project and it’s going to be an issue.”
Sood also highlights the need for top management to be actively involved for BI initiatives to work, emphasizing that BI implementation is not just an IT function.
“I’ve seen three back-to-back failures with three different providers,” he says. “If your BI initiative is completely driven by your IT guys, it is not going to succeed.” The end-users must be involved to make sure the project is not just a nice piece of technology. It should be aligned with business objectives.
BI Is Not for Everyone
The good news is that it is easier these days to guard against expensive failure. Compared to the past where “you would spend the money [on infrastructure] and then look for a use case,” says Sood, cloud computing has made possible rapid prototyping and building of the business case before actually building infrastructure.
But he warns that just because it is easier to implement BI initiatives does not mean it is for everyone.
“All BI initiatives can be classified into three areas,” says Sood. “It helps increase revenues, reduces costs, or measures improvements in services rendered.”
“If your BI analytics are not in any of these three buckets, you’re wasting your money.”
Chief Analytics Officer Europe
15% off with code 7WDCAO17
Chief Analytics Officer Spring 2017
15% off with code MP15
Big Data and Analytics for Healthcare Philadelphia
$200 off with code DATA200
10% off with code 7WDATASMX
Data Science Congress 2017
20% off with code 7wdata_DSC2017