The time of the "hunters and gatherers" has passed; big data is getting into the details. "Our objective is to mine the collected data to find answers to the challenges faced by the respective company," says Dr. Thomas Erwin, Global Execution Partner Data & Analytics at KPMG. The consulting firm teamed up with IT industry association BITKOM to assess the current status of big data use at German companies. The result: increasing numbers are basing their business decisions on data analytics and are succeeding in applying the findings profitably – notably in the machine tools, plant engineering and automotive industries. This means innovative data analytics becoming increasingly important for business decisions.
Production planning and project management benefit
According to the study, the vast majority of those surveyed analyze their customer data: 79 percent analyze sensor or site data and 70 percent use publicly available data, such as general information on economic developments. The insights gained are applied to production planning and project management, as well as customer analysis. Finance and controlling are important application areas: 91 percent of companies use data analytics in risk management, to identify and assess potentially disruptive developments. In the insurance industry, for example, these involve risks in specific risk groups; in the construction business, the risk analysis is often performed during the bid phase. Companies use it to determine whether they might have underestimated the financial scope to continue in order not to submit bids that are too cheap.
Axel Oppermann, Avispador Analyst
The increasing use of such methods isn't entirely uncontroversial, however, as some companies seem to emphasize collection above all else. Bitkom figures say one-third (34 percent) of companies overall have a big data strategy. There are major differences between industries, however: 56 percent of media companies and 46 percent of insurance companies have such strategies, while the corresponding figure in the automotive sector is just 34 percent. As such, some companies still have difficulties finding the right strategic approach to big data – as Axel Oppermann, an analyst at Avispador, reports (see interview).
How the right big data strategy works
BARC, an analyst firm, knows that this strategy can be designed with a practical orientation. Accordingly, companies should clarify the general conditions first, before launching big data projects, specifically management support as well as the potential for generating innovation from the data. The next logical step is twofold: identifying and prioritizing use cases and implementing a closer examination of the data. This means examining the following questions: Which data has not yet been sufficiently combined and analyzed? Which data sources, internal or external, could deliver additional value? Will the data enable implementation of the potential use cases at all? Put another way, the actual work starts long before the "hunt" for data.
According to the Bitkom study, specialist expertise is needed to ensure that big data projects don't founder – in both the strategic development and the subsequent execution. After all, when information from different sources is combined, the success of analytics projects often lies in the variety of the data collected. "To achieve it, companies have to work together with data professionals who are specialized in complex analytical methods," advises the association. And last but not least, the user companies have to take a proactive approach to subjects like data privacy and IT security. Otherwise, big data can quickly turn into "big chaos".
Big Data… three questions for Axel Oppermann, Avispador
Mr. Oppermann, as an analyst, where do you still see challenges in the use of big data?
Challenges still include understanding the concept at all and estimating its potential, but the primary issue is key personnel. Moreover, some users do not think in terms of opportunities – that is, sensible, value-generating solutions – but instead concentrate on facts, figures and data.
A good example of this – or rather, a poor one – was the marketing manager of a mid-sized retailer we advised. He didn't know the difference between correlation and causation. As a result, he compared a huge amount of data and identified a correlation. Although the time series were similar, they had nothing to do with one another; there was no causation involved. Instead, a different variable turned out to be decisive. His bad luck: he made major decisions based on this incorrect interpretation.
…and when you talk about opportunities, what do you mean?
When used properly, big data can be a component for modernizing management approaches and business models, in turn creating the foundation for ensuring our economic status quo. An example of this is assuring the quality of healthcare provided by analyzing DRGs: diagnosis-related groups in medicine.
In your opinion, what do we have to do to seize such opportunities?
We mustn't draw false conclusions, especially when the results cannot be interpreted or when the "wrong" questions are asked before the analysis. To analyze the data correctly, we need high-quality data, qualified resources and – above all – intelligence. Without the necessary intelligence, the data is practically useless.