For something so vital to the long-term success of a modern business, the concept of business intelligence is not well defined. But that doesn't stop many companies from wanting it, even if they don't entirely understand it. Here we'll take a look at this IT business trend, what it is and how it works to improve a company's processes.

What Is Business Intelligence?

Business intelligence (BI) refers to the collection and analysis of data in order to produce insights that will improve a company’s processes. There is a lot packed into that definition and, as a result, a lot of the confusion around BI stems from the assumption that it stops with analysis. Although the distinction gets muddy sometimes, business intelligence can be thought of as the end goal of business analytics because it produces the actionable insights a business needs to make informed decisions. In order to do this, effective business intelligence needs to meet four major criteria:

  1. Accuracy
    This refers to the accuracy of the data inputs as well as outputs. The two are, of course, related. Any system that requires analysis can fall prey to the garbage in, garbage out (GIGO) problem, in which tainted data can ruin results, even when the analytical model is sound. In order to get accurate answers (output), the data going in must be accurate and relevant to the questions the business is seeking to answer.

    It is often impractical to try to dump all the data produced by a company into an analytical model and expect it to make sense of everything from production numbers to employees' marital status. This is why human discretion is often used to select the data that is relevant to a particular problem. That said, this selection can be over-exercised or simply done wrong, bringing us back to the GIGO problem.


  2. Valuable Insights
    Not all insights are valuable. Knowing the handedness (left or right) of the majority of your customers may be useful for a baseball glove manufacturer, but would be of less use to a shoe manufacturer. Although crunching all the data to find out something that was previously unknown can be satisfying, BI should offer concrete insights. For example, if analysis showed a sports store that many customers who purchased baseball gloves also purchased running shoes, the owner could rearrange the store displays to cluster shoes and gloves for customer convenience, or separate them to different corners of the store to maximize the chances of browsing.


  3. Timeliness
    Getting accurate and valuable insight is only half the battle. Business intelligence must also be able to deliver those insights at the right time. If the aforementioned sports store only discovers the glove and running shoe correlation in December rather than at the start of the buying trend, it may lose the opportunity to capitalize on that information.

    There are two parts to timeliness - the timeliness of the data going in and the timeliness of the insights coming out. Businesses have different decision time frames depending on what they do. A retail outlet will likely want to be feeding very timely sales information into BI with the hope of getting timely insights to be implemented on a monthly, weekly or even daily basis. Longer-term operations like an oil and gas exploration and production company may only be interested in insights on a quarterly or yearly basis.


  4. Actionable
    The final hurdle for any type of business intelligence is to provide insights that can be acted upon. To some extent, this means gaining an understanding of practical constraints. For example, virtually any company could become more efficient if it had unlimited capital to upgrade of all its equipment. So, good business intelligence should identify the upgrade that will produce the most return or, better yet, other utilization schemes that would make the most of existing assets. In other words, business intelligence should provide insight beyond what is obvious and work within a company's unique constraints to deliver actionable ideas designed to improve a business's processes and, ultimately, its profitability.

The BI Process

So what exactly is being done in the black box of business intelligence? The business intelligence process is very similar to the Deming cycle. It has four broad steps that loop over and over (the buzzword for this is continuous improvement, or Kaizen).

  1. Data Gathering: Data sources are identified, and the data is collected and converted into a format that can be analyzed.
  2. Analysis and Action: The data is analyzed and a course of action is taken.
  3. Measurement: The results of the action are measured using a chosen model.
  4. Feedback: The results of the action are used as another data point to make ongoing improvements to the BI process.

Business Intelligence in Action

BI is a Deming cycle applied across an organization and all of its business lines. It is usually facilitated by technology. In this view, the software merely helps make this process much easier to implement and allows for a larger sample of data to be included in the analysis. At the end of the day, however, BI is only effective if it is trusted and used to guide human decisions. That said, the leaps BI has made in guiding large organizations has helped give it a considerable amount of credibility in the world of business. This means many companies want BI - even if they don’t entirely understand it.