Logistics chains consist of a large number of partners and their services. Today, organisations have to wade through enormous amounts of data to evaluate the efficiency of their processes. Key figures become the starting point for improvements and greater economic efficiency. The faster they can be analysed, the more room for manoeuvre companies gain.
Correlations become clear from looking at past periods. Only real-time analyses, on the other hand, offer the ability to specifically control day-to-day business. What both views have in common: Business intelligence (BI) systems produce reports with better data in much less time than manual data research, which is often collected in Excel spreadsheets.
If you want to make good use of a BI system, you should answer the following questions first:
The core task of BI systems is to give users regular reports based on consistent criteria. This enables them to identify trends and make comparisons over longer periods of time. Data can also be explored, i.e. viewed in depth through drilldowns. The resulting insights can then be used to make strategic and operational changes.
Freight forwarders, for example, can determine the efficiency of their local transport fleet: How many stops does a delivery vehicle make per trip? What is the average loading or unloading time per stop? How many minutes does a truck spend in a traffic jam during its trip every day? Statistical reports let managers see whether they are operating profitably. If not, the reports suggest where and what to improve. For example, a driver could complete a trip in reverse order to avoid traffic jams during morning rush hour.
Shippers, on the other hand, are more interested in the lead time of their shipping orders, i.e. the time from order entry to loading on the truck: How long does it take on average for a picker to start the order? How many pick positions does a picker have to process per order? How long does picking take on average? How long does it take to book the transport and prepare the shipping documents for each order? – Key figures that provide insights into warehouse setup and the efficiency of the commercial shipping processes.
For analyses to be accurate and meaningful, they have to obtain their data from a reliable source. It is best to combine information from all relevant production systems in a linked database. This becomes the single point of truth – the place that always holds the most current and reliable data. This is the basic requirement for real-time analyses and prevents fuzzy interpretations. If the data was collected manually at different times, the analytical result will be distorted. Up to now, it has been common practice in many freight forwarding agencies and other medium-sized companies to compile monthly reports from Excel spreadsheets. That poses a serious risk for the meaningfulness of the results.
That is why the ideal approach to setting up and maintaining a BI system not only calls for a central database, but also requires it to be filled regularly. As quickly as possible. This is only way to generate analyses in near-real time, which means the data is suitable for managing the day-to-day business.
Before company managers give new instructions to their employees, they have to clearly understand what the data analysis is telling them. The systems provide a variety of visualisations. They can then be used to generate report templates. Bar and pie charts or line graphs with sparklines, however, still have to be interpreted. There are now software systems that translate data into text for this purpose. These language generators use templates for producing a formulated description in seconds (natural language generation). This is done using the numbers and rules associated with them. For the reader of a report, a steep downward line thus becomes an "exceptionally negative end-of-month closing". This makes the reports easier to understand for recipients with different preferences.
Depending on the viewer's perspective, the data can have entirely different meanings, too. A logistics service provider, for example, will draw different conclusions than its customers. It is even possible to grant customers access to BI information as long as the system is configured based on rights (as a single point of truth). They only see those parts of the analyses that apply to them: for example, all the shipments in a given period and their delivery rate. The service provider's trade secrets remain hidden.
Still curious?
Add a comment