With the abundance of live data available in social media, asset management, works delivery and too many more to list its worth a few minutes to think about what is classed as real time and where it’s appropriate to use real time data.
I recently sat through a presentation for a planning and reporting system where the words ‘near real time’ were mentioned a few times around the display of dashboards and reports. Being a person who naturally likes to know all the facts, the word ‘near’ got me thinking about what was actually classed as ‘real time’ and how this affects the way in which we report and or interpret data. Having real time data may be a benefit in certain situations, however is this all that we need or is there something missing?
This led me to thinking about the answer the question of
“Where do we draw the line between Operational reporting and Analytical Reporting?”
Broadly speaking data and information used in reporting can be used in two ways depending upon the type of reporting or dashboards required. Let me explain the difference between these two types of application to give a better understanding in the differences.
For me operational reports / dashboards sit at the heart of what defines real time reporting. These reports are based upon dynamically changing data, sourced from live systems or processes. The information displayed in these reports is used to make the ‘here, now and tomorrow’ decisions. Therefore depending upon the process or system the data could be live or a snapshot taken at frequent intervals; such as minutes or hours apart. The frequency of the snapshots would depend upon the frequency of updates made to the system or the needs of the business. For example if the system was only updated daily there would be little point in taking snapshots every 5 minutes. Operational reports are good for front line staff and line managers to take rapid decisions on the course course of action. Much like the dashboard in a car. The information displayed there is needed to make decisions now. For example how fast am I going, how much fuel do I have left, what is the temperature of my oil?
Analytical / tactical reports and dashboards however are based upon historical snapshots of live data, looking at measuring or analysing the volumes of data over time. The big difference therefore in analytical / tactical reports is the ability to look at trends and changes in data over past days, weeks, months and years… However in many cases this will also require the building of supporting data marts and transformation processes to process and house the snapshot data taken. The business will need to define the metrics and trends required as generally transactional data extracted from operational systems will need to be transformed to calculate new metrics and enable subsequent data analysis. By reviewing the historical performance this data is therefore used to develop trends and models to make decisions and predictions on future performance or outcomes. Using the car dashboard example above, for analytical / tactical reports I might be interested in the average speed over the duration of my last journey, or the average mpg, or maximum oil temperature. In this case I’d need to snapshot the current speed, fuel consumption and current oil temperature at regular intervals to then analyse the data and determine the averages from the data.
Moving back to my original thoughts about ‘near real time’. I believe for any Business Intelligence deployment a line should be drawn between real time and near real time availability along with the distinction between operational and tactical / analytical reporting. BI Mangers should help the business to decide on the type of reporting needed and how the information is to be used to determine data refresh frequencies, metrics for analysis and reporting type. With these simple answers in place the right reporting solutions can be developed and avoid confusion and complications in producing meaningful Business Intelligence.