By Michael O'Dwyer, Contributor
Big data technology is allowing more companies to collect vast amounts of information about their customers. But figuring out what to do with that data while protecting clients’ privacy is becoming a big issue for many businesses.
For that reason, companies need to understand their goals before investing in big data, industry experts say.
“Many companies have big data, or are starting the process of collecting data without understanding what eventually they want to achieve,” says Ohad Hecht, managing director at emarsys APAC, a Hong Kong-based global provider of automated e-marketing solutions with headquarters in Vienna. “What [do] they want to learn about their business, which data sets should be connected in order to translate it into information for decision makers and on what frequency should it be delivered?”
There is no denying the practical benefits of big data, which can determine patterns of consumer behavior and help companies adapt to changing markets. Many businesses already are using advanced software to extract relevant data automatically.
Understand our business better’
“We are using big data to understand our business better: from our business [key performance indicator], clients’ behavior, and interactions with our company,” says Hecht. “We combine [customer relationship management] data, financial data, HR data, billing and bank data into a single big data warehouse and visualize it in different dashboards that target information needed for decision-making in different areas of the business.”
But some experts think that you need human analysis to complement what computers can do.
“Companies, even large players, have issues with analysis and prognosis because of its high complexity,” says Yvonne Hofstetter, managing director of Teramark Technologies GmbH, a Germany-based provider of big data technologies and artificial intelligence for the industrial Internet. “It requires mathematicians and [data scientists].”
Hofstetter believes that the term “big data” is a buzzword created by marketers, as her company does not distinguish between raw data and “smart” data.
“The core of big data, as we understand it, is neither the storage nor the retrieval of raw data,” she says. “It is the analysis of big data and the inference, which is provided by artificial intelligence (AI). This AI can be of any combination of swarm intelligence [a technique where computing resources/sources of raw data are utilized from multiple sites or remote locations], expert systems or optimization.”
Gathering external data
Analyzing data from external sources is often a challenge, especially in the supply chain, which is usually outside a company’s control.
“In more complex scenarios such as global trade, business transactions take longer and are more complex,” says Greg Johnsen, chief marketing officer at GT Nexus, an Oakland, Calif.-based operator of a global cloud-based platform for international trade and supply chain management. “Slice-in-time snapshots won’t suffice; they won’t tell the whole story. Transactions include and rely on a broad multilevel network of separate organizations.”
There is no single source of data for this type of network, Johnsen adds. According to Aberdeen Research, 80 percent of this data resides beyond the four walls of the business in the supply chain.
“Capturing and harnessing this data is a major challenge,” says Johnsen.
More work is needed before companies can effectively utilize the benefits for B2B commerce.
The opportunity for big data in the world of global commerce is enormous and powerful but remains elusive,” Johnsen concludes. “Only through radically different technology and network-centric information models designed for operational engagement between companies can the potential and promise of big data be realized.”
Ethical issues of data gathering
The other challenge facing companies is protecting customers’ privacy.
A recent Revolution Analytics survey indicated that 88 percent of 865 data scientists believed that consumers should be concerned with the data privacy issues involved in data collection. In addition, 80 percent believed that an ethical framework is necessary.
Teramark’s Hofstetter agrees with the findings.
“Personal data does not only mean name, address, phone number,” she says. “It’s all the unstructured data that, if being assembled, form the digital twin’ of a person. And because that is the case, big data has an ethical perspective, as it is tied to a person, which is completely overlooked by big data proponents.”
“You cannot simply do what you want with this data,” she adds. “The owner of the data is the person who produced it. Analyzing, sharing, selling is simply unacceptable if the data owner of the data does not approve.”
Hofstetter advises companies to use human analysis.
“Employ mathematicians,” she says. “Do not rely on trainees or programmers for data analysis, like German companies do. Big data analytics is much more complex than you might think. It requires experience with data, and their dependencies, and with mathematical models and AI.”
Hecht disagrees, saying computerized analysis is better.
“Many companies see big data as an area that belongs to data scientists’,” he says. “In fact, the analysis should be executed automatically. Analysis by itself is a means to an end. The information must be actionable so companies can make use of their investment in big data.”