IMO there is no general right or wrong in the approaches else it will depend on various factors which method is more suitable for the scenario.
Important is to know that Qlik isn't just a front-end tool else it has also enormous strengths within the ETL part and could be seriously regarded as a dwh which just used a direct file-storage without a container-overlay like most of the db or dwh uses.
This must not mandatory be a real disadvantage especially not in regard of the easiness and speed to develop datamodels and solutions and most notably not in regard of the performance because by applying a multi-tier data-architecture with multi-level incremental approaches it would be quite hard for a dwh to get near to the performances. dwh's may have benefits in respect of integrated backup's, version-control, multi-user working, access-control and further governance and administration aspects but sometimes those features are so heavy that they aren't (properly) used and nearly everything could be also implemented in Qlik.
We have rather a small environment in regard to the used datasets and the report-side but we do everything in Qlik because the company dwh doesn't contained all needed data, it's quite difficult and lengthy to get the appropriate data (multi-level ticketing) and it's much too slow. Further our demands to evaluate and display the data require very expensive ETL on often changing business requirements which we couldn't realize with a dwh. We need solutions from today to tomorrow and not in the next year …
- Marcus
