Back in 2001, with the release of the now cult-classic film Donnie Darko, there was a website that came out to accompany the film. The site was (and is) as cryptic and vague as the movie. There is this quasi linear path through the site but at the same time no clearly delineated objective - you experience the site and you do so just for the sake of experiencing it. It doesn't need to tell you what theaters the movie was being released in, or when it was being released because you could get that information other places. This site was for enthusiasts, for die-hard fans. It was something to play with that added to the richness of the film by extending the experience of the film.
Now contrast this with a site that sells movie tickets. There is a clear objective. You use it to buy tickets online to save you from having to buy tickets in person. I go to these sites with a purpose, I complete my task, I leave. It's a tool to do a job.
The balance between creating a tool vs. creating a toy isn't always this clear. You want to create a pleasant experience, even something cool, but if the primary objective is to enable people to complete a task then you should focus on building a useful tool. Help your users do the thing they came here to do. In the case of QlikView this means building applications that focus on the user's needs, apps that clearly show the data, that help users explore their data and find new insights.
I've seen more than a few applications where people, with the best intentions, tried to create a fun environment but ended up adding too many pieces of unnecessary visual "flare." When considering a background, a chart, an animated logo, or a photograph you need to ask yourself: "is this going to help people use this application?" If the answer is no, then perhaps it is best left out. There are plenty of well designed applications that take a very minimalistic approach to design. The application should be well designed, but well designed this isn't synonymous with having a trendy aesthetic. You don't need to cram in all the design tricks you know.
Real design solves problems and the application you are building should be a tool to help people solve problems and complete tasks.
