Corporations of all sizes spend a significant amount of money on Business Intelligence software (BI). The business intelligence category is nearly $20B annually, according to Gartner Group industry analysts. And it seems there is plenty of tail wind to this sector for the foreseeable future.
Like many business software categories, there is more promise in the sizzle than in the steak. Even at its most basic, what’s terrific about business intelligence tools is that they can visually display many different data sources in easy to digest charts and graphs. And these visual aids can easily be accessed by users wherever they may be. Further, the user can pick and choose certain data sets to be overlaid in the same chart or graph in a few clicks.
The irony with Business Intelligence is that it really doesn’t have any “intelligence” pre-built into the system. There are two main themes that make “out of the box” Business Intelligence tools rather Unintelligent: One is data, and the other is knowledge.
For nearly all Business Intelligence offerings “out of the box,” it simply displays what you ask, drawing from whatever data sources you decide. The issue with most companies is not the simplicity of visual display, but rather the lack of data quality. So, if the underlying data is not clean, accurate, and up to date, neither is the accuracy of your requested outcome. According to a recent Experian report, 33 percent of organizations globally are planning a data cleansing project in the next 12 months to help improve business processes such as selling and marketing. And what causes most ‘dirty’ data? Humans. So, if you are left manually trying to integrate all the data and hygiene, it can get really ugly really quickly. You may think “Well, I don’t have that much data,” but think about all the different data sets that could exist in one singular customer view of the selling process in the Outdoor Power Equipment industry. It ranges from product registrations, to POS data, to pricing MSRP vs. actual retail sales, to prospects, to customers, hundreds or thousands of product SKUs – all necessary, all needed in one combined view.
Knowledge is intelligence. Software needs humans with experience to understand the nuances of the proper output to build algorithms (fancy word for mathematical equations) before the data displays to a user. Otherwise they may not know what to make of the data, or they could be trusting the output as is and making decisions on dirty data at best, or inaccurate models at worse. Think about the recent models for the last two United States Presidential elections. Candidate Mitt Romney’s ORCA analytics team missed the boat on the model, and yes, Mrs. Clinton’s Ada team did as well. And the one who got it right, yes, Mr. Trump’s team. Politics aside, the point is made. The modeling is the key to determining the outcome. The software just processes the data that humans (experienced or not) tell it to.
So before you invest too much on decisions that can effect big profits or losses, or software (costing a mere fraction of total upside), think about what you need to win and the supporting model that needs to be set-up. Then monitor it. Challenge it. Tweak it. Acknowledge it. Optimize it. It takes a team of dedicated and diverse technical experts with street-wise business know-how who focus on this as their core competency. And trust me, this team could care less about the colors of the chart, and more about the more accurate outcome of the effort. Just ask Clinton or Romney.