Business intelligence (BI) is broadly understood to be the process of examining the mountains of corporate data every company stores and coming up with pearls of wisdom that can be used to improve business processes and ultimately the bottom line.
Sifting through data to find trends that can assist in the management decision-making process is a far better way of running a business than simply going on gut feel. However, as many companies have discovered, parsing historical information has a limited benefit as it can deliver solutions that are too broad and generalised, and sometimes too far after the fact to help management change the business for the better.
"BI as we know it has delivered value to business for years, but competitive operating environments are forcing companies to rely on new predictive analysis applications to enhance their BI systems to find commonalities, trends and unique associations in customer data, allowing them to refine their selling efforts," says Charl Barnard, GM of business intelligence at Knowledge Integration Dynamics (KID).
"Predictive analysis is not a long-term project that delivers results once every quarter or once a month when a data warehouse is updated," Barnard adds. "It happens in real-time to enable companies to take immediate advantage of whatever is happening in the market instead of reacting after the fact.
"In other words, predictive analysis has changed BI, making applications less about storing and rehashing data and more about creating real knowledge."
Going beyond assumptions
Traditional business intelligence starts out with an assumption. Data is analysed with a thought or potential outcome in mind and the results obtained are all anchored to the original assumption. Business leaders with the experience to make educated guesses can often make assumptions that provide veritable pearls of wisdom they can incorporate into their strategies. However, their plans are still limited to their own knowledge and experience.
Predictive analysis goes further, beyond human assumptions to uncover information that would not have been known otherwise. The software has the ability to recognise patterns and associations between seemingly disparate bits of data, irrespective of any ideas users may have. Information supplied by predictive analysis assists in refining marketing efforts on the most profitable target markets in the areas most likely to produce the best sales. It can also assist in product development, with information, for example, on what colours or designs sell better in particular areas.
"Traditional BI processes would not find this information unless users knew it was somewhere in their data and specifically constructed queries to pull the information out," adds Barnard. "Predictive analytics sorts through data and finds these links, and then predicts what is likely to happen if management takes a particular course of action. Decision-makers can now devise different action plans on the back of this information and test each to find which delivers the best results for the business."
More than numbers
One could be forgiven for believing that predictive analysis, according to the above description, is simply BI with a new name and a few extra features. But predictive analysis goes further and can also identify patterns in soft, non-quantifiable data. For example, customer comments can be scanned for patterns of denigrating words to determine locations where customer service is below par, or if certain parts of products, manufactured in the same factory are regularly reported as faulty.
In addition, predictive analysis tools do not deliver reams of data for decision-makers to work through. More usable and accessible formats such as charts and graphical representations make it simpler to identify and act on information. Even non-technical users can identify trends and items of interest at a glance and drill down by simply clicking on a particular bit of information.
"Once again, management can test hypotheses and determine the potential outcomes of strategic decisions with ease," Barnard notes. "Once the best strategy has been determined, they can run cost comparisons and ROI queries to predetermine the full scope and outcome of their actions before spending a cent. And they can do it in less time than ever before."
Predictive analysis takes the magnifying glass that was BI and turns it into a self-guided microscope programmed to find and highlight information organisations didn`t know was there. "There is an enormous amount to learn," concludes Barnard, "and with predictive analysis you don`t even have to know where to start looking."
Share
Editorial contacts