|
but also an element of experimentation,” said Pain.
While reporting after the fact remains an important business imperative, the ability to make decisions in real time wasn’ t supported. While it was great for the C-suite, it didn’ t help line managers who wanted to be able to make more immediate decisions. As a result, line managers began creating their own systems in Excel or using desktop database applications. The trouble was that different parts of the business could use the same data and, by applying different analysis techniques and their own definitions, end up producing inconsistent results.
|
© iStockphoto. com / Evgeny Terentev |
||
Agile BI systems make it possible to |
|||||
6 |
deliver the data more consistently, in real time and in a form that allows line managers to make real- or neartime decisions with confidence. Rather than working with raw data and then manipulating it manually, they are able to work with data sets that are better tailored to their needs and delivered in a more timely manner.
Collins clarified this saying,“ All data has a cause and an effect. It’ s not just about reporting the data, it’ s about trying to link traditional metrics to corporate performance.”
We are already seeing this in retail where companies can see what you’ ve browsed and purchased online, use demographic data they’ ve gathered and then target specific advertising or product recommendations.
The way this works requires a shift in how businesses store and manage data. In the past, most of the organisation’ s data was held in a structured database and then distributed outwards to the
|
data warehouse. In today’ s world, we’ re looking at massive volumes of data, from multiple sources, and much of it is unstructured. According to Harrison, this means were more likely to see a Hadoop-style cluster as the main data repository with in-memory processing systems used to extract and manipulate the data for analysis.
However, the data modelling that used to happen at the start of the analysis process is performed after the data enters the database, not before. Harrison calls this“ just-in-time modelling”. Businesses that aren’ t able to leverage the unstructured data are“ just going to lose” in his view.
It’ s tempting to see the issues around using data as being largely technical. However,“ You can change and manipulate data until the cows come home but unless you have a direct connect to the business that’ s all you do,” added Collins.
The good news for the CIO is that many of the skills needed to support the business in using business intelligence
|
tools already exist in the IT department. However, they’ re often channelled towards a different purchase. Using live data from systems to monitor and manage performance is something many systems managers have been doing for some time.
Applying similar techniques, we’ ve seen companies such as Sportsbet, Coles and iiNet use BI-like tools from Splunk to monitor their systems in real time and manage performance. For example, by monitoring server performance during peak periods, they are able to reassign resources dynamically to ensure that systems not only remain available but perform at optimal levels.
How information comes into a business and is used to deliver benefit to the bottom line is critical. Grant Christian, from Information Builders, says,“ The information delivery chain is much more than just delivery or presentation of information, it extends out into the cloud with social analytics in one direction, but then also extends right back to the
|