Business Impact of Information Delay
Separate data management platforms for systems of record and for systems of decision are the norm today. While transactional applications need data to be organized in a way that helps serve business transactions, analytic applications need the data to be organized in ways that help serve business intelligence purposes, such as historical trend analysis. Historically, it has been too difficult to optimize a disk-based database to serve both purposes at once, so separate database architectures emerged — one for transactions and one for analytics.
Because data in a separate analytics platform is usually not current, it isn’t possible to run analyses on such data that can usefully inform decisions made in the midst of a business process. The information is only as current as the last time the data was moved to the analytics platform and transformed for analytic usage. The inherent latency of the data inhibits the business in the following ways:
The speed of business processes is confined to the performance characteristics of the transactional database, which, in turn, is dependent on the arcane tuning of its disk-based operations. If some analytics are introduced into the process, then the performance of the analytical database and of the data movement software can further slow down business processes.
If analytics data is used in interactions with customers and suppliers, such interactions can be complex and cumbersome as sales, support, and other staff must switch awkwardly between their transactional applications and the analytic applications that provide relevant reports and visualizations. The ideal solution would be to integrate real-time analytics into the transactional application, but this is problematic when the transactional and analytical data is kept in different places.
For example, a contact-center agent would have to look away from the transactional application during an interaction and try to find relevant analytics. Even if this could be done, it's likely that the analysis will have been based on data gathered prior to the current interaction episode.
Without the integration of analytics into transactional applications, there is little opportunity to build into such applications the flexibility and dynamism to adjust to shifting business circumstances. Building applications that can optimize business operations, or recognize cross-sell and up-sell opportunities, is extremely difficult and business innovation is inhibited.
With the complexity, inefficiency, and inflexibility of current data systems, any analytic activities that pertain to in-flight business processes are largely unavailable, especially those that involve complex analysis. This is illustrated in Figure 2.
Well over 40% of IT polled said that users could not use predictive analysis or simulations or work with real-time data because of these limitations. Big Data workloads and diving down into more granular data than is routinely available were also off the table for almost as many respondents.