Ferrari and Russo are among the best writers on tabular models anywhere, but this is not their good book: it is an introductory pamphlet Microsoft gave away to push a new product. Much of the text is 2016 screenshots, personal gateways and content packs, a concept the platform itself has since retired.
What I took from it
The chapter on building the model. That is where the distinction between a measure and a calculated column lives, which sounds minor and is what separates a report that answers instantly from one that crawls. A column is computed row by row and stored; a measure is computed when someone asks, against whatever is filtered at that moment. People who are fuzzy on that end up storing things they should be computing.
Out of it came the first question I ask when someone wants a dashboard: at what grain does each table you want to combine sit. A report that puts sales by invoice next to budgets by month produces numbers that do not reconcile, and the worst of it is that they fail silently, with no error on screen. The budget integration section of the book is exactly that case, and it is the only moment when the text stops being a tour of menus.
The second thing I took is less pleasant: most dashboard requests are not a visualisation problem, they are a symptom that nobody has agreed on how a sale is counted. No tool fixes that.
What I left out
Everything else, without regret. The screenshots, the content packs, the personal gateway, the query editor as it looked then. Anyone wanting to learn from these two authors should go to what they wrote about the calculation language and about modelling, not to this welcome booklet.