SSRS and the Tabular server/Vertipaq engine/DAX are probably the two most valuable components, leaving out SQL Server (and Azure versions of it, including the Azure warehouse and big data technologies). SSRS's display functionality is resoundingly flexible, and by comparison to other vendors, relatively easy to learn. SSAS Tabular performs very well with modest effort in design, and the ability to perform complex business modeling through DAX calculations stands out in a crowd.
Microsoft BI really entails a whole catalog of products. The relative ease with which these products work together is where the primary value comes into play. When you buy SQL Server Enterprise, you automatically have access to SSAS, which can handle Tabular or multidimensional cubes and data mining, SSIS offering a pretty comprehensive integration toolset, SSRS, covering just about any paginated and subscription-based reporting, and of course all of the in-built features of SQL Server, the individual useful features of which there are too many to spout out here. Each of these components work well enough in isolation, or as part of a larger ecosystem with other vendors, but together are a well-integrated machine that can handle end-to-end needs on the entire lifecycle of data. Being one of the more popular vendors, Microsoft products are also the first compatibility for other vendors' products. The breadth of products do a fine job at enabling the delivery of data and insights to the right people, in the right format, at the right time. Doing so efficiently/quickly is where a huge amount of value comes in. Many other enterprise vendors are capable, but are either slower to execute, more work to maintain, cost much more to find talent and skills for, or some combination of all of the above.
No product is perfect, and the Microsoft stack is no exception. The first things that come to mind are Microsoft's recent shift in strategy to focus on cloud first. While this makes sense for them, it leaves the on-premise products lagging in support and new features, while the vast majority of users (mid to large corporations) are still using on-premise solutions and/or are cloud-averse. For example, the announcement that the pricing model for Power BI would change, simultaneously rolling out both Power BI Premium and Power BI Report Server created a great deal of confusion for people who were very enthusiastic about the product. The failures to address concerns with the speed and nature of the changes were felt by users across the spectrum, and it was a situation created by their cloud-first strategy. Similarly, if you look at the tried and true on-premise options for data processing, interfacing with emerging technologies, especially big data technologies, lags behind. This is considerably less true for Azure users, but that is specifically a cloud offering, which again, many corporate entities are not yet ready to embrace.
It's easy to think of BI as only the visualization aspect of data, and that point of contact between users and data is absolutely where the rubber meets the road in BI. In reality, there are a whole stack of tools and concepts behind the visualization that enable that interaction - from security, data governance, integration, performance, network and infrastructure, and automation... there are many facets to BI as a system. The visualization aspect, while being the most visible to business users, also happens the be the weakest point in the entire Microsoft BI stack. The PowerBI visualization experience is underwhelming in almost every way compared to many alternatives.
Another difficulty is navigating the many products within the suite. There are many components, and each component has many versions. Dealing with feature and compatibility issues with so many versions of so many products can be very frustrating at times. Microsoft does not do itself any favors on this front with the way they name their products. "PowerBI", for example, could be in reference to the desktop design tool, a cloud-based service for publishing and administering data models (which comes with three distinctively different pricing models, the features of which are different and not interchangable), or an on-premise server solution replacing SSRS. You'll make yourself dizzy looking at the Azure services offered. The good news is the sky's the limit, the bad news is you'll have to navigate some pretty cloudy areas to make heads or tales of what to actually use.
Four years, using largely SQL Server 2012 and related versions/components.
In general, no. Most stability issues encountered have been more related to network or infrastructure, and not the products being used.
As with any system, good design is paramount. I have certainly run into times when performance suffers, or changes and enhancements are slow and expensive. In every case, redesign has solved the issues. As hardware improves, and the many options (especially in the Azure and big data space) continue to evolve, designing for scale has become easier and easier. One of the difficulties that naturally occurs in this ecosystem is simply knowing how to design with so many options that can potentially do the same or similar task.
I have experienced mixed results on this front. We have had some very available and supportive account reps at times, and other times not so much. My biggest complaint would have to be that the disorganization from Microsoft's side makes the results on this front inconsistent.
Depends on which components we are talking about. For reference, however, I don't think there is a single Microsoft component that was as painful as any given Oracle component to work with.
Do your homework when deciding what components to leverage. The worst thing you can do is try to use them all. The gamut of products under the Microsoft BI banner enables every form of BI - choose the ones that serve your specific purposes, and leave the rest on the table until a new need arises.
