The exponential growth of Big Data- large collection of data sets that can be analysed to improve decision making- highlights its increasing relevance to all economic stakeholders, though it may come at moral and legal risks. Nevertheless, I believe the value created can create mutually beneficial solutions.
Improvements to the type of data collected or a different approach to data analysis is becoming a significant competitive advantage for companies. For instance, exploiting the volume of high frequency nature of data collection enables retrospective metrics such as consumer sentiment to be ‘nowcasted’, that is, calculated and tested in real time. Company decisions based on such information will improve operating margins. For example, institutional investors who possess the computing capabilities to perform Big Data analyses on consumer product reviews are able to determine stock price directions following the introduction of a new product rather than use traditional sale figures. According to Gerard Tellis and Seshadri Tirunillai, the more negative reviews, the bigger the likelihood of a stock price drop, though positive news lacks a statistically significant impact. Portfolios based on their findings outperformed the market by up to 8% on an annualised basis. I believe such data analysis can be extended to produce inferences regarding investor behavioural psychology. Given the asymmetric effects on stock prices from positive vs. negative reviews, it highlights the risk averse nature of investors in general. Nevertheless, it may also mean positive information from marketing campaigns have already been fully reflected in the stock price.
Clearly, the biggest challenge is ensuring timely exploitation. Timely exploitation is vital, as highlighted by the success of the institutional investors’ model portfolio. Companies, particularly those with diverse business lines, must ensure data is shared between these departments. Financial institutions who fail to share data from their financial markets to their lending or money management lines are preventing themselves from developing a more comprehensive and coherent understanding of relationships and trends in customers or financial markets.
Big Data projects require high infrastructure costs to support the real time capture of such large quantity of data points. Incompatibility of formats and legacy systems challenges effective data integration and analytics application, emphasising the need to create a technology stack through analytical software apps. However, the unclear return on investment (ROI) hinders guaranteed funding. Through actual Big Data projects, core business figures can be used to measure ROI, though pitching a business case with an accurate ROI initially is difficult which slows adoption. Research implies only 25% of CIOs are prepared to launch tech agendas and Big Data IT expenditure for 2015 is forecasted to consist of only 22% of IT spending. I believe companies should not narrow their focus on building their up front ROI, instead find scalable, low cost infrastructure such as the Cloud to configure and store their data architecture. The Cloud Dashboard’s tracking of KPI metrics from multiple databases can boost business transparency and therefore improves decision making.
Even when new infrastructure is set up, the need to interconnect a complex number of data sources also slows work. CIOs can also adopt an incremental approach to the business case- instead of pooling their entire data source into one data warehouse, they should find value in using data of individual data sources- an ‘analytics approach’. In identifying key data in each data source, CIOs can avoid the risk of generating noise and allow specialised infrastructure to maintain the data they were built to.
When data channel amalgamation is relatively easy to implement and benefits override its costs, an organisation can magnify its potential from Big Data. This is because the more data available, the more patterns can be recognised and therefore predictive analysis is increasingly useful. The US healthcare system’s data channels include clinical, claims, pharmaceutical research and patient records, each presently managed by individual constituencies. However if they implemented reforms to integrate the databases to analyse costs and health outcomes of medical treatments, physicians could be guided to the optimal treatment in terms of cost and outcome, thereby boosting productivity of the health sector. In fact, President Obama is currently developing a Precision Medicine Initiative, where genomic sequencing of patients is to be made accessible to researchers, who are then able to develop a new generation of personalised medicine especially targeting genetic variants.
However, acceleration of biomedical research and development requires public support for the sharing of patient information. Steps to ensure greater transparency over the use of medical data and incentives such as criminal penalties for data misuse will improve public trust. Greater transparency will demonstrate the benefits of these initiatives- though this does not address the risk that the sheer volume of data from different data sets increases potential identification of anonymous patients which may lead to employer discrimination against individuals’ medical history. This highlights the need to implement consent declaration, despite this eroding the overarching unbiased data accumulation feature of Big Data analysis.
Others argue the need to balance the socio economic value generated by Big Data with social values of equity, fairness and privacy still circumvents the bigger picture. Quantification is an important tool though improper use of data is worse than having no use of data. Yahoo’s performance reviews of employees (ranked from 1 to 5) aimed to motivate though had opposite effects- talented employees no longer wanted to be a team, and employees lacked motivation to be mobile as this increased the possibility of a lower score. This demonstrates quantification boosts bureaucratic power and overly values short term impacts due to their ease of measurement. To ensure quantitative techniques remain objective in their analysis, judgments must be made on information that balances context, incentives and other factors.
Nonetheless, despite the condemnation of data collection techniques due to its apparent moral hazards, the ultimate solution is not to restrict the selectivity criteria of data to be collected, but to educate how data can be fairly used. Almost all large organisations have real time data warehouses, which, combined, contains more up to date information on our economy than all our government agencies. A public-private sector partnership will enable the government to use such private sector data warehouses to create more knowledgeable decisions regarding fiscal and monetary policy. In fact, Google encourages government regulators to stop deleting google searches as using their trends, they were able to track the flu outbreak even before health agencies could. Thus, condemning data is condemning knowledge. Take pharmaceuticals as an example- their reticence to social media platforms means they are failing to exploit a Big Data opportunity, as they are ignoring patient information shared online. If pharmaceutical companies employed data scientists to mine such qualitative real-world data, it could add to their information sources on drug side effects and their effectiveness. Now is never a better time to benefit from Big Data, particularly due to the availability of tech vendors who automate social media analytics such as Treato, Liquid Grids and IMS (branched into social media in 2014 as a result of their acquisition of Semantelli), who each minimise the burden on companies’ central objectives.
Ultimately, Big Data must be developed in a system which addresses matters such as public-private data sharing agreements, frequency of meta data updates, transparency and technical standards. That way, value can be created not only for commercial companies, but for the public sector and collective societal well-being.
Financial Times, Ivey Business Journal, Time, Wall Street Journal , CIO, BioTech IT