Business Intelligence Goes Beyond the Hype in 2013
2013 isn’t the first time business intelligence has been a top priority, but it could be the year it will actually be implemented.
The biggest challenge (in a BI implementation) is setting the right expectation. People interpret BI as reports. That is the first disconnect, as with that interpretation comes the interpretation of granularity. People interpret reports at the highest granularity. BI cannot be at the highest granularity, it has to be arrived at the highest granularity.Veneeth Purushotamanbusiness head, technology & supply chain, HyperCity Retail
In today’s fast-paced digital age, given the amount and the rate of data generation, it is incumbent on enterprises to use data to gain insights and stay one step ahead of the competition. Fortunately, CIOs are stepping up to the challenge. According to the State of the CIO Survey, almost 90 percent of Indian CIOs are at varying states of implementing—or are in the process of implementing—a business intelligence solution. Out of that, 56 percent say that they expect to complete their business intelligence and analytics projects within the next 12 months, a sure sign that CIOs realize the power—and the need for—business intelligence.
However, just because BI seems to the in-thing, it doesn’t mean that CIOs will have an easy time getting the necessary funds or support from management to drive the project.
This was a lesson Meenakshi Agrawal, VP-IT, Mumbai International Airport, learnt. When she proposed a business intelligence solution to optimize the airport’s operations by monitoring key performance metrics—a project that is currently underway—she had to wait six months for the necessary approvals to come though.
“It was quite a tough battle,” she says. “Unless you’ve seen it or worked with it (BI) before; it was more theory for users based on screenshots from vendors. They were not able to touch or feel the benefits. The process was protracted and it took a long time to take the decision to move ahead. They were uncertain about whether the solution would be as beneficial as planned.”
Part of the challenge is that business intelligence and analytics projects are seen as expensive, that’s what 49 percent of CIOs report in the State of the CIO Survey. And while many think of a BI project as a sure way to drive business, another 30 percent of CIOs say that one of the key barriers to BI is that it has no clear ROI.
It helps for stakeholders to be able to see how the solution will work to get them on to invest in BI. This is what Agrawal did. “We had to have a number of demos by vendors. We asked them to take our data and put it into their tool and allow users to see their own data being manipulated. In this way stakeholders felt more comfortable,” says Agrawal.
Nilesh Sangoi, CTO, Meru Cab Company, who is also looking to optimize his operations using data analytics as a tool—he plans to set up an algorithm that will automate and increase the efficiency of the booking process—went about getting buy-in using another way. He says to get approvals for 20 mini-projects, for which analytics was used extensively, he ensured that for each of them proof of concepts were carried out by running controlled experiments and demonstrating the benefits from the projects. This, he says, helped build greater acceptability in rolling out bigger implementation across the board.
The Report Mentality
Another thing that CIOs who want to implement business intelligence should be prepared for is expectation management. And that leads to another challenge: The confusion over what BI really is. “Most of the time, BI is misinterpreted as MIS. BI is not just reports, BI is beyond reports,” says Veneeth Purushotaman, business head, technology & supply chain, HyperCity Retail.
“The biggest challenge (in a BI implementation) is setting the right expectation. People interpret BI as reports. That is the first disconnect, as with that interpretation comes the interpretation of granularity. People interpret reports at the highest granularity. BI cannot be at the highest granularity, it has to be arrived at the highest granularity,” says Purushothaman.
What he means is that the performance of a BI solution will degrade, as the granularity of the information that is requested increases, leading to a point where performance could fall below a user’s expectations—if they aren’t expecting it.
The reporting mentality is not easy to get away from given how entrenched Excel is within most organizations. According to CIO research, the majority of Indian organizations, 74 percent, use spreadsheets or Excel to share BI or analytical insights.
What is it Good For?
Trying to gain a deeper insight into data can be quite problematic for a number of reasons. First, there are the various departments within an enterprise, each with its siloed data. So, when making reports the MIS team will have to work through Excel and this leads to a situation where, as Purushotaman puts it, “ there are multiple versions of the truth, none that tally with another.”
In Agrawal’s case, the request to get a report on a certain category of information would be done as and when requested, in an adhoc manner. “We didn’t have much historical MIS analysis available on the fly. For example, there was no easy way for management to know the rate at which cargo imports/exports were growing or trends in passenger loads. Normally the MIS team worked manually with Excel sheets and prepared graphs as required. This is the background to us creating a dashboard for senior management with all the data coming together in ways that were easy to slice and dice,” says Agrawal.
For Purushotaman, the fact that a lot of reports were Excel-based, which raise security concerns because they could be copied, was an issue that was identified for resolution by opting for BI solution.
Another benefit of using BI is that it gives users “the freedom to make whatever report they want,” says Purushotaman. But the biggest benefit of using BI, according to Purushotaman, is that “by having a data warehouse you get to put out one version of the truth; a single version of truth that everyone sees. When you have MIS there are Excel sheets sitting at very employees system and there are a million versions of the same data.”
Agrawal also looks to benefit from “enabling a one customer view” of all the required information, wherein different officials can then see all the necessary information on a “dashboard without having to request for separate reports (from various departments),” she says.
Raring to Go
The performance of a business intelligence solution is very important factor to consider when explaining to stakeholders how a system will actually work. “Make stakeholders understand there is a lag and it takes time for information to be shown. Keep them informed very clearly on data availability and performance,” says Purushotaman.
When such steps are not taken, you can end up with a situation where stakeholders decide that the system is too clunky and call for its termination, a situation Agrawal almost found herself in.
When the first tests were being carried out, MIAL users found that information was being fetched in too slow a manner for their liking. In fact, they wanted to close the project on account of this performance level, says Agrawal. But luckily she dodged this particular bullet, and while they are still facing a few performance-related challenges, Agrawal is hopeful that they will move ahead according to plan.
Despite these challenges, there’s no denying that Indian enterprises and their CIOs are determined to own the benefits that come with business intelligence. 2013 could be the year of business intelligence.
Eric Ernest is correspondent. Send feedback to firstname.lastname@example.org
CISOs have been hearing for some time now that they need to learn how to "speak the language of business" better. It is one way to gain respect and avoid being viewed mainly as a scapegoat.
When the Cloud Security Alliance (CSA) launched in 2008, the questions around cloud computing centered around whether cloud was secure enough to be trusted, how it could be managed securely and in such a way as to keep regulators happy. There was also plenty of talk about whether cloud would fully displace traditional enterprise data centers.
“Data scientists are not necessarily a must for every organization. Data engineers who execute most of big data projects are more crucial than data scientists,” says Ben Sadeghi, Data Scientist, APAC and Japan, MapR Technologies.