Power BI's agility drops as the complexity of the analysis and the volume of data grow. That happens in scenarios where the organisation's expectations are not met. In many cases these impressions come from an inadequate comparison between the organisation's needs and Power BI's capabilities before the implementation is carried out. This article sets out to explain the main problems and solutions when those scenarios come up, which is common in organisations starting to use this kind of low-cost application.
While Power BI is an excellent tool that lets non-technical users connect to data and begin exploring and analysing it in an orderly, fast way, it should not be forgotten that the objective of these BI solutions is to speed up operations by reducing the time spent processing and accessing information. If no improvement is found in those two aspects, then the solution is not adding value to the organisation.
The most visible effect of reduced performance shows up the moment you start using the tool, because it consumes considerable resources simply by being open. While the installation can be done quickly and without problems, it may be that the machine is not the right one. If the machine — a laptop or otherwise — is not adequate, the impact will be apparent immediately, because it will run slowly while doing other things such as checking email, accessing web platforms or using the company's applications. That is why we recommend taking advantage of the application's free licence and testing its performance in a real working scenario where multiple tasks are carried out in parallel. That exercise will produce information about the capacity of the resources the organisation has and whether they can support this class of application.
If your organisation passes that test, it is in a position to begin more rigorous ones, which consist of evaluating the application's performance in carrying out operations that are commonly done with macros, pivot tables and other Excel functions. Power BI has two methods for connecting to data sources. The first is called Import, which copies the data from its sources to the analyst's machine. This connection method is recommended when there is little data, because it is stored in the machine's memory, which means reduced performance. The second method, called Direct Query, exists to free up the machine's resources, because it copies only the data that will be used in the views defined. However, it is in this scenario that new needs are created, because on its own this alternative has limitations.
Among the main needs that arise is the use of the DAX and MDX query languages, which require a slightly more technical profile, as well as a level of aggregation — consolidation — at the database layer to speed up queries and procedures. And it is precisely from this point that you begin to define whether it is viable to go for a Power BI–type solution or for other, more robust and scalable alternatives on the market.
Later articles will go deeper into some concepts and solutions for various scenarios and requirements associated with Business Intelligence and Business Analytics solutions.
