A plea for reducing complexity in reporting

What's more, with “EXACTLY our workflow” we are lying to ourselves. There is no such thing as ONE workflow. Everyone operates a little differently, depending on the day, project and mood. We work differently on Mondays than on Fridays, customer appointments change routines, home office as well. In short, it just makes a lot of sense that you don't have to reinvent the wheel for everyday processes.

And lo and behold: the software we introduced does what it's supposed to! Suddenly we have reliable time tracking, transparency about absences and overtime and an illustrious selection of setting options that, with a small dose of pragmatism, cover all the special curls we think we need to afford. It's wonderful!

But where there's light, there's also shadow (and yes, 5 bucks into the piggy bank for using phrases): Our purchased standard software fails at the evaluation level. The interface initially appears pleasantly reduced: Five menu items, one called “Evaluations”. Sounds straightforward, everything is fine, everything is wonderful. But as soon as you open this item, the classic disadvantage of standard software suddenly becomes apparent: In the exaggerated eagerness to please everyone, you please no one. Instead of serving me the two or three really relevant reports on a silver platter, I get lost in a jungle of options.

Common problems when confronted with a multitude of predefined evaluation options

• It's simply too much: Each report may be useful in its own right. But we don't just use one tool, we use many. And if each tool comes with 27 reports, it just gets too wild. I want to focus on reports and quickly gather insights, but instead I get stuck in the maze of variations.
• It's too confusing: The one view that I need is there - but I can't even grasp it due to the multitude of selection options or I regularly forget where or under which name I can find it.
• It is not differentiated clearly enough: What is the difference between a “working time history” and a “ history of working hours”? Why is the “Working hours overview” in the “Overtime and absences” category, but not in the “Times” category?
• It is too similar: If there are too many evaluations, they will inevitably end up being very similar at some point - the marginal utility decreases and becomes a total loss.
• It is too specific: Many reports pre-filter data differently. If I compare two reports, I have to check exactly whether they are based on a comparable database. This takes time, requires concentration and is anything but intuitive.
• It's not my evaluation: If I have a specific question, I usually want to think for myself about which data view I need to answer it. If I only have standard reports to arrive at an answer, I need to know exactly whether there is a suitable view and which one it was.

The underlying data model for our time recording is not particularly complex: employee comes in, takes a break, continues working, leaves, may be on vacation, may be sick. Familiar to everyone, not rocket science. So why do we need 27 different views? I don't need special views for every conceivable question - I need a few, powerful tools that answer my questions flexibly. It must remain usable! I'd rather make ten intuitive clicks in the right tool than theoretically arrive at my target image with just one click, but have to study the software for years to do so.

How to enable complex evaluations without being confusing? - A few things that have worked for us:

• DIY dashboards: Let users build their own dashboards, where they decide which selection of standard reports they want to see regularly - and which they don't
• Create views: Let users save their filters to access them on a recurring basis without having to reconfigure each time (or check whether the desired setting is still cached correctly)
• Intuitive report names: Let users assign their own names that are intuitive for them, not for the manufacturer.
• Make tables editable: Let users show and hide columns, sort, adjust column orders.
• Set priorities: Give different relevance different presence: reports that are expected to be used more frequently should be more present than those that “might be nice to have once in a while”
• Form clusters: Bundle similar reports sensibly, not arbitrarily.
• Less is more: If in doubt, make your existing report configurable instead of putting a very similar report next to it.
• A matter of taste, but I'm a fan of it: Starting from the general data set: Build reports so that they function at the highest possible level and therefore use as few clearly identifiable pre-filters as possible. If you want to go deeper or cluster, use additional filters. I don't trust a report that uses magic and coincidence to determine which data it contains. It must be plausible which data records are included in the report and which are not.
• A data anchor: One of my big problems with standard reports is that they all use some sort of pre-filter and I never know which records will slip through. Perhaps they only slip through because certain properties are not maintained or are maintained incorrectly.
• Global filters: There are sometimes use cases in which I don't want to configure each evaluation individually, but instead want to set overarching filters, for example, so that I don't have to filter each evaluation individually.

My conclusion

It's good that we develop individual software and that I can easily make speeches without 200 customers with 300 opinions breathing down my neck. But seriously: there are certain basic principles of UI design and clarity is one of the top priorities. Very few software solutions are primarily used by power users, who can be expected to accept a little complexity. For the masses, I don't need an equally large number of solutions, but rather a few, intuitively usable tools that are simple but powerful and, above all, transparent with regard to the underlying database in my software.