Business intelligence must encompass both traditional BI and new innovation.
As the technology behind business intelligence and analytics continues to evolve, it seems every year there are new buzzwords hyped as crucial to any type of BI project success.
, data governance, big data, Internet of Things. I hear them a lot at the Gartner conferences that I attend each year to learn from the leaders of our industry, and I bring those lessons and inspiration home with me to TARGIT.
BI jargon aside, the success or failure of any BI project ultimately depends the accessibility of critical data to those who need it – and only those who need it. This has historically been a problem for many analytics projects. This guide will help you bring analytics to users in every department and increase user adoption rates.
BI must cater to the dual demands of IT and BI business users. A truly data-driven organization has set data free to decision-makers throughout the organization. But robust data governance is needed to ensure data is only ever seen by those who are meant to see it and that the data is clean, accurate, and up-to-date.
Slick dashboards and user friendly data visualizations are incredibly important, but a business cannot run on dashboards alone. In the past, ease-of-use has not necessarily correlated with data accuracy and consistency. The traditional data warehouse and business intelligence practice is in conflict with the experimentation and innovation of big data and analytics. Gartner compares this to suits versus t-shirts.
In order to become a fully data-driven organization, you need both. What Gartner calls a “bimodal” BI solution tackles both the necessity of agile, user-friendly analytics and reliability and security of data. A bimodal BI strategy should not only facilitate traditional business operation—the classic data warehouse and continuous decision loops—but also discovery and innovation.
Bimodal BI is both centralized (company-wide initiatives) and decentralized (change, innovation, and exploration). Traditionally, these business intelligence disciplines were seen through the lens of “either/or.” Either/or is a thing of the past, according to Gartner. They think that by 2017, 75% of organizations will be employing bimodal BI, up from 45% today.
Indeed, in order to be successful with big data analytics, experimentation is a must. The ability to play with big data sets and analyze them on top of what’s already in the data warehouse encourages employees to think strategically without the need to pull in IT. We encourage this to happen in smaller, select groups, which we call sandbox analytics. But I’ll get to this in my next post.
TARGIT has an array of customers employing this strategy already with TARGIT Data Service, such as drive-thru dairy and grocery service Swiss Farms. For years Swiss Farms had valuable data stored in disparate spreadsheets on desktops throughout the company. The company now easily pulls in Excel and other flat files on top of data from the data warehouse with the click of the mouse.
Check out Data Service in work:
Today, Swiss Farms is using TARGIT Data Service to gather informative social media data sets and mashes that data up with traditional data warehouse anlyses. Swiss Farms IT Director Chris Gray harnesses Twitter and Facebook data from customers and plans to mash it up that data with internal customer call center data to gain a more comprehensive understanding of customer satisfaction. With the right inforamtion at hand, Swiss Farms creates a better customer experience for their customers.
To reap the true benefits of bimodal BI at your own organization, company culture must be adapted along with a technological platform that caters to both modes. Often though, taking on that task of analyzing big data can be a big, intimidating projects for companies.
Learning how to work with big data starts by first gaining an understanding of the smaller data sets you have within your own organization. You’re likely already picking up a lot of data sets that resemble big data in nature, but are limited in size. This “small data,” as I’ll call it, is an excellent starting point before venturing into big data projects. To get familiar with big data approaches, I would normally recommend the following approach:
Get to know the data set. Understand the nature of the data you are looking at, so that you can find a meaningful way to summarize it. Fast prototype iterations with tabular tools like TARGIT Data Service enable you to understand the correlations of the data.
One of the big challenges with big data is mapping it to your company’s situation. In other words, how can you put it into a useful perspective that will help you reap the benefits of advanced analytics?
Again, start by correlating the “small data” you worked with in step one. Correlation points can include be time, place, events, etc. Explore different correlations and combine them. This is a time for experimentation. With multidimensional models, you can easily explore your many options. This step will reveal insights that can carry on to the next step, which is …
Don’t keep your findings a secret. Working with this type of data might not be 100% bulletproof in terms of data quality, but there is likely a good story being told. Therefore be sure that you can share your findings easily. Feedback from your peers might enable you to think differently about these new data discoveries or see your findings in a new perspective. Regardless of where you go, failure is the first step towards success, so don’t be afraid to challenge things.
Don’t confine yourself to a world of 50%. Embrace the 100% of bimodal BI and experience the difference in your business strategy.