Tag Archives: Predictive Data

Evaluating big data and predictive analytics

bigdata

The use of business intelligence and marketing information systems has expanded in recent years. Through advancements in technologies, marketers can extract value from very large data sets. Very often, companies can benefit if they use and reuse the same data to extract added value from it. Sometimes, it would also make sense for these companies to acquire data that they do not own (or data that was not collected).

All individuals leave a “digital trail” of data as they move about in the virtual and physical worlds. This phenomenon is called, “data exhaust”. Initially, this term was used to describe how Amazon.com used predictive analytics in order to suggest items for its customers. Predictive analytics could quantify the likelihood that a particular person will do something — whether it is defaulting on a loan, upgrading to a higher level of cable service or seeking another job. Such data anticipates human behaviours that have not happened as yet. For instance, Fedex has predicted which customers were most likely to defect to competitors. Even, Hewlett-Packard made a good use of suitable data to identify employees that were on the brink to leave the company. The latter corporation took remedial decisions in anticipation of staff turnover.

Predictive data is usually based on large amounts of cur¬rent and past indicative information that may have been collected from multiple sources. Such data could also provide additional details of customer personas, segments and prospects. Quantitative techniques can be deployed to find valuable patterns in data, enabling companies to predict the likely behaviour of customers, employees and others. First Tennessee Bank had used predictive analytics to increase its marketing response rate by better targeting its offers to high-value customers (IBM, 2015). Through predictive analytics businesses’ could quantify how many consumers will buy their products after receiving electronic mail. They may also measure how effective their personal mailing was.

Nowadays there are fewer inaccuracies in the measurement of big data. In addition, many applications of data can arise far from the purposes for which the data was originally intended. However, big data and predictive analytics could raise a number of concerns. Minor increases in the data accuracy of predictions can often lead substantial savings in the long term. There many companies that have saved significant financial resources by using predictive analytics. For instance, “Chickasaw Nation has used predictive and patron analytics to reduce its month-end close processes by 50%. This way it has also improved customer experience. In a similar vein, predictive tools and smart cards enabled Singapore Land Transit Authority to provide a more convenient transportation system.

Although, individuals tend to regularly repeat their habitual behaviours, pre¬dictive analytics cannot determine when and why they may decide to change their future preferences. The possibility of “one off” events must never be discounted. Many customers may be wary of giving their data due to privacy issues. The underlying question is; when does personalisation become an issue of consumer protection? In 2012, consumers learned that Target was using quantitative methods to predict which customers were pregnant. Very often, advances in technology are faster than legislation and its deployment. These issues could advance economic and privacy concerns that regulators will find themselves hard-pressed to ignore. It may appear that digital market manipulation is pushing the limits of consumer protection law.

Evidently, society has built up a body of rules that are aimed to protect personal information. Another contentious issue is figuring out the value of data and its worth in monetary terms. In the past, companies could have struggled to determine the value of their business; including patents, trade secrets and other intellectual property.

Despite its numerous pitfalls, the market is responding to the emerging demands for corporate IT solutions. Extant relational databases are capable of handling a wide variety of big data sources. Statistical analytical packages are similarly evolving and are working in conjunction with these new data platforms, data types and algorithms. Furthermore, big data is also being modified for those clients that may require cloud-based services. Cloud-based service providers offer on-demand pricing with a fast reconfiguration facility.

This short contribution suggests that in the foreseeable future many corporations would require bespoken software that is relevant for their particular line of business. Customised business intelligence software and big data systems allow organisations to load, store and query massive data sets in short time periods. Business could make good use of structured data (such as demographics) and unstructured information (including text and images) to improve their operational performance and customer service levels.

Advertisements

Leave a comment

Filed under Analytics, Big Data, Marketing

Crunching Big Data for Operations Management

Big data

For decades businesses have been using data in some way or another to improve their operations. For instance, an IT software could support small enterprises in their customer-facing processes. Alternatively, large corporations may possess complex systems that monitor and detect any changes in consumer sentiment towards brands.

Recently, many industry leaders, including McKinsey, IBM and SAS among others have released relevant studies on big data. It transpires that they are using similar terminology to describe big data as a “situation where the volume, velocity and variety of data exceed an organisation’s ability to use that data for accurate and timely decision-making” (SAS). These providers of business intelligence solutions have developed technical approaches to storing and managing enormous volumes of new data.

The handling and untangling of such data requires advanced and unique storage, management, analysis and visualisation technologies. The terms of “big data” and “analytics” are increasingly being used to describe data sets and analytical techniques in applications ranging from sensor to social media. Usually, big data analytics are dependent on extensive storage capacity and quick processing power requiring a flexible grid that can be reconfigured for different needs. For instance, streaming analytics process big data in real time during events to improve their outcome.

Insightful data could easily be retrieved from the Web, social media content and video data among other content. Notwithstanding, such data could be presented in different forms; ranging from recorded vocal content (e.g. call centre voice data) or it can even be genomic and proteomic data that is derived from biological research and medicine.
Big data is often used to describe the latest advances in technologies and architectures. Nowadays, big data and marketing information systems predict customer purchase decisions. This data could indicate which products or services customers buy, where and what they eat, where and when they go on vacation, how much they buy, and the like.

Giant retailers such as Tesco or Sainsbury every single day receive long-range weather forecasts to work 8-10 days ahead. Evidently, the weather affects the shopping behaviour of customers. For example, hot and cold weather can lead to the sales of certain products. It may appear that weather forecasting dictates store placement, ordering and supply (and demand) logistics for supermarket chains. Other retailers like Walmart and Kohl’s also use big data to tailor product selections and determine the timing of price markdowns.

Shipping companies, like U.P.S. are mining data on truck delivery times and traffic patterns in order to fine-tune their routing. This way the business will become more efficient and incur less operational costs. Therefore, big data extracts value by capturing, discovering and analysing very large volumes of data in an economic and expeditious way. This has inevitably led to a significant reduction in the cost of keeping data.

Big data can also be linked with production applications and timely operational processes that enable continuous improvements. Credit card companies are a good illustration of this dynamic as direct marketing groups at credit card companies create models to select the most likely customer prospects from a large data warehouse. Previously, the process of data extraction, preparation and analysis took weeks to prepare and organise. Eventually, these companies realised that there was a quicker way to carry out the same task. In fact, they created a “ready-to-market” database and system that allowed their marketers to analyse, select and issue offers in a single day. Therefore, this case indicates that businesses became much more effective (and efficient) in their processes through iterations and monitoring of websites and call-centre activities. They could also make personalised offers to customers in milliseconds as they kept tracking responses over time.

Organisations are increasingly realising the utility of data that could bring value through continuous improvements in their operations. This contribution indicated that relevant data needs to be captured, filtered and analysed. Big data is already swamping traditional networks, storage arrays and relational database platforms. The increased pervasiveness of digital and mobile activity, particularly from e-commerce and social media is leading to the dissemination of meaningful data – that is being created each and every second. Successful, online businesses can gain a competitive advantage if they are capable of gathering and crunching data.

Leave a comment

Filed under Analytics, Big Data, ICT