What are the benefits of data warehouses and OLAP from a business and information systems perspective?

Question

OLAP

Hello.
Books on OLAP sound the standard sets of theses that justify OLAP / HD.
For example:
- quick analysis (no more than 5 seconds for processing).
- provision of logical/statistical analysis.
- multidimensional representation of data (+hierarchies).
- You can access data regardless of its size.
etc.
But how, in simple terms, can you define business goals that are clearly declared for business? How to evaluate them? Let's say in terms of finance? Or perhaps there are non-financial dimensions? For example, the speed of processing tables or generating reports (was-will be)? Judging by the periodicals, the maximum that the companies received was: "the time for creating management reports was reduced." What about numbers?
So, the original task:
- a chain of stores creates its own new accounting system.
- the system is global, unites terminals in warehouses and stores into a single network.
- it is planned to collect statistics and information into an operational relational normalized database (OLTP).
- and build reporting - according to the standard scheme - by generating calls to the operational database.
And specific questions:
- Why create another database (HD), albeit somewhat modified?
- Why create a special database (HD), if it is enough to schedule a copy of the existing OLTP database?
- What do you say against the idea: we don't need HD. What for? Having made a simple copy, we will not load the main database with reports.
- Why buy a BI tool - if programmers write a series of reports to an existing OLTP database?
- What can BI+HD give that a set of reports cannot give to accounting on 1C and to a clear and simple relational OLTP base of a scientific system?
- Can a purchased BI tool simplify the development of an accounting system and, in fact, simplify / speed up the development of a reporting building block? (Is it possible to bring reporting into the "BI + HD" layer?)
I would like to hear answers that would indicate the benefits of business in a quantitative and measurable way. Or examples from your experience.

Answer 1

In your case, it is really not necessary to invest in BI from the first day. A replicated copy of OLTP and reporting built on it is enough. As my experience shows, only after the business itself understands that this reporting is not enough, it will be possible to talk about the implementation of analytics.
No need to build storage and BI for the sake of BI itself (although sometimes they do it - image, increasing the value of the enterprise, etc.). You should first try to close business problems with the available reporting development tools and only then think about storage technologies.

Answer 2

Thank you. Regarding "the business itself will understand that this reporting is not enough" - I will note. that businesses are already formulating various reporting requests. We counted more than 28 reports on initiatives. And - not related to accounting, only in the subject area - goods, logistics, etc. If it goes at such a pace, we will pass exactly over a hundred reports.
Again, we return to the thought - isn't it easier to give the business "cubes" - what would they build and analyze themselves?

Answer 3

the structure of the HD is usually a "star" for each of the tasks.
In this structure, it is much easier to formulate queries than in OTLP with many relationships. So a business user is unlikely to understand the names of OLTP tables and fields, while there are almost no problems in writing queries to the data mart. The data is already prepared, cleared. The structure of the data warehouse is completely query-oriented (in OLTP, the structures are focused on inserting and changing data). The presence of a separate data warehouse and OLAP, as a rule, completely removes reporting issues, most of them can be built by business users themselves (Ad-Hoc) - the most useful and frequently used ones can already be transferred to the list of "built-in"