BI trends in SA

There are several business factors driving key BI trends. Insights, which will help businesses understand more about their customers and transactions in order to improve business processes and enable 'next best action' or 'next best offer' recommendations.

Secondly, actionable decisioning, or the move from people interpreting information subjectively and driving some actions, to objective, fact-based real-time decisioning across the enterprise.

In addition, customer expectations play a factor. Omni-channel, seamless real-time interactive communications and services is expected by customers. "The end goal is to use data and information better, to optimise our enterprise decisioning in order to close the gap between current metrics and the optimum business model."

So said Julian Ardagh, CEO of Effective Intelligence, during his keynote address at the ITWeb Business Intelligence Summit 2014 in Bryanston this morning.

He said there are also several data factors driving key BI trends. "Data is a competitive asset whose value is growing exponentially, and is coming into the organisation from internal controllable sources as well as external sources. BI tools have typically been developed to report on relational databases with well-structured data standards. However, the quality of data in terms of suitable content within a business context is still a major issue."

In addition, he says the vast volumes of unstructured data from external sources, such as social media, require new methods to turn this data into usable intelligence, which has resulted in an explosion of non-relational databases and new data management methods. "There is a need for CIOs and CTOs to understand the array of new technologies and BI tools to manage these needs and deliver better intelligence to meet the business objectives. A balance must be established between the BI tools needed for internal structured data, and large unstructured data sources."

Ardagh says BI has evolved to assist knowledge workers to derive insights from patterns in business data that can be converted into actionable insights in order to improve business strategy, tactics and processes. "Currently, BI is in a phase where powerful interactive visualisation and query abilities are mandatory, along with better mobility to enable speedier action. BI tools are evolving to be simpler and more intuitive, to enable a wider group of users to take advantage of them."

The first trend he cited is BI versus predictive analytics. "BI is about looking for relatively basic patterns in data that will offer easily communicated visual insights that allow a user to better conceptualise the problems and apply appropriate actions to improve either processes or thinking."

On the other hand, predictive analytics, he added, is about using powerful statistical and mathematical algorithms to identify complex patterns in data that can then be used to build statistical models whose output can determine the likelihood of one or more future actions or events; it's more about solving business problems.

"Predictive decisioning is the integration of multiple predictive models into decision management platforms to enable interactive, automated processes."

Secondly, he said analytics outsourcing is growing. "Companies, even with strong analytics in one or more departments, are starting to outsource certain analytics where they need to. This helps them gain benefits instantly, particularly considering the current economy."

Another trend that Ardach sees is dynamic pricing, using predictive models. "Dynamic pricing attempts to optimise defined outcomes based on consumer behaviour. It assesses the demand, supply and risk factors to find the balance for each service interaction, and price accordingly. It only works where consumers recognise that behaviour changes affect risk or reward."

Data quality and enhancement solutions are another trend spotted by Ardach. "Data quality has been a major inhibitor to effective BI queries for some time now. With the trend towards customer-centricity and householding, the requirement to fix more address data faster, means that many companies are either licensing software that does the whole job without the need of any domain specialisation within the company, or they are outsourcing the problem to specialists."

He added that many businesses are requesting enhanced information be attached to their customers to better segmentation and communication targeting or are outsourcing this as well. "The trend is to do it now, before POPI comes in next year.

"Big data is another trend, and tackling unstructured data drives the need for new methods and tools. A traditional keyword search is inadequate; it must be based on context such as the semantic context. BI searches must be capable of content deduplication and ambiguity resolution, such as inconsistent references to the same person, thing or process in multiple sources. Ultimately, it must improve decisioning, not just provide insights."