Many clients, especially when starting their QlikView journey, face the decision of when to use larger consolidated documents and when to use separated, domain-specific documents. Having built the first document it is tempting to just keep adding to it. There is no definitive right or wrong answer to this dilemma. Instead it comes down to an evaluation of the relative costs and benefits.
There are lots of things to consider, but the final decision will come down to the balance of:
- requirements for information
- flexibility to adapt to new and changing requirements
- performance implications
- licensing costs
Because QlikView is so good at what it does, it becomes natural to want to provide all information to all users. In reality the various information consumer roles have quite different needs.
At one end there are the operational roles where there are a high number of decisions being made. Operations’ need for information will be quite detailed, often at the transactional record level and will be narrow in its domain and scope. Contrast that with strategic roles where the number of decisions is much lower but each decision has high impact on the organisation. The information needs for strategic roles is quite different with broad domains being brought together at a summary level.
Having everything in one document will work when the application and information needs are mature and stable. However separate documents are better placed to cope with evolving needs and sudden business environment changes. Separate documents are also much simpler to build, maintain, reconcile.
A single large document won’t perform as well as separate documents where (a) users are only accessing summary totals which need to be calculated on the fly as opposed to pre calculated and (b) where the server needs to allocate more RAM due to the size of the document. The RAM issue arises because the server will typically allocate approximately 10% of the document per active user. A large document for all users exaggerates this issue as each user ends up with more RAM than necessary and there are more users! A single document also makes it harder to adopt a qvd (QlikView data file) based data load strategy, thereby losing out on benefits such as incremental loads.
Smaller purpose built documents can be expensive, however, if they are repeating the same information multiple times for different uses.
A single large document will appear to reduce the license requirements as users only need one CAL to get all their information. This may not be as pronounced as expected though because a clever design will have most users only needing to access one document and only a small number needing to access multiple documents. The cost for those people is effectively capped as using a named license provides access to unlimited documents.
What you may notice from the discussion above is that there is very little in the way of pure technical considerations. For example clicking on a URL to open another document with pre-applied filters is much the same as clicking on another tab and section access can replace document-level security.
Whatever your organisation has decided to adopt as its strategy around the consolidated versus multiple document decision, make sure it has done so for the right reasons. Use the criteria above to assess all the costs and benefits involved and don’t be locked into an initial choice that may have become outdated.
You can read more about Professional Advantage and QlikView here.