Writen by Stefan Farrugia • 28th September 2017
One of the most common problems I encounter when meeting businesses is their ability to make good use of the high volumes and variety of data that they generate. Any system is enshrined to make use of its data, secure it and understand it. However, the ability to integrate other external data streams is insufficient. There are two issues which effect even highly capable operational systems. Firstly, these systems typically have a weak reporting function. Secondly, even if it the reporting function is sufficient, the data is still limited to what is contained within the system itself.
When a business has multiple systems, data analysts export static information and then compile, merge and aggregate data into an external tool like Excel. This not only entails a laborious process but often produces reports that are static, outdated and prone to errors.
So, in an ecosystem of disparate systems and Excel sheets, a business intelligence (BI) solution provides the ability to disrupt the process by freeing up time for users who should be identifying challenges and opportunities instead of wasting hours gathering data.
Implementing a BI solution reduces the dependency and the time taken to generate reports. It automates the repetitive tasks through well-defined data warehousing principles ensuring correctness of data. In this way, data is processed, transformed and moved into a reporting structure which is easy and fast to query.
The most important achievement is the completeness of the data sets which are made up of multiple sources. The tool also has the ability to mash data and expose new valuable insights through a multitude of reporting options. It is imperative to understand that there is a fundamental challenge that needs to be mitigated once all systems start talking to each other. This is consistency!
Data is known to be complete within a system. However, when data from different systems are combined, some inconsistencies arise. At the first stages, users will be required to react and correct as many of these inconsistencies as possible. Once the data is cleaned up – ideally setting rules and validations in the sources – any new inconsistencies will be exposed immediately through the BI solution.
When the data warehouse corresponds to the single version of the truth, then one can start adding layers of intelligence, such as predictive analytics. I always advise that before becoming too excited about predictive analytics, businesses have a golden opportunity to analyse historical data, understand the mistakes and achievements, and learn what needs to be repeated or corrected in the future. At this stage, KPIs are defined, grouped and consolidated. It's also important not to keep KPIs at executive board level but share them across the business. The more people are able to use data to improve their contribution to the business, the higher the overall achievement.
The impact of such a strategy on the data security should also be assessed carefully. Usually, data elements are protected and secured into their systems. However, through a data warehouse, they become exposed completely and if the solution does not cater for data element security than there is a very high risk that people might be able to access restricted information ignoring the need of clearance.
Finally, but very importantly, it is imperative to make sure that the solution is built around the business needs and not around the systems in place. When systems are upgraded, the BI solution needs just a connection change and some reconfiguration, and not a complete overhaul. When the business model is implemented irrespective of the underlying systems, users have more flexibility and power to change the systems when required.
Having business data at users' fingertips is no doubt beneficial to any business. Data warehousing is a solution for combining data from different sources. It ensures data correctness and enables business to have more time to analyse data rather than gathering it. Reports can be generated rapidly and shared across the organisation to empower multiple stakeholders to assess performance data and take decisions faster on further performance improvements.