I came across a post which discussed a list of must-haves for identifying information on printed reports. I came up with my own to see how they matched. All this sounds really simple, but too often the basics are missing.
1. Report title
2. Data source
3. Data refresh date/time
4. Print date/time
5. Page number (of n pages)
6. Filter values, if any
Now, what did Mitch Wheat have?
He adds:
- who printed it
- alternate shading of data lines
- don't print empty reports
- URL to the live report
Apart from those, we're substantially in agreement. Personally I don't think those additions are must-haves - although they're useful options. He adds to the title a "unique identifier" which he rightly points out is essential when the stock of reports is large.
What else makes you grind your teeth in frustration when it's lacking in the report? (apart from the quality or absence of data!)
Too often, there's a communication disconnect when someone responsible for BI services is working through issues with information users. It's not always possible to even agree whether they are working off the same page. Are all six identifiers above present? If the software tools aren't able to generate all those identifiers, it's time to rethink the toolset. If this is not possible, proxies (through either data or business processes) need to be engaged. For example: strict controls over when data is refreshed; information users only having access to a single data source. In other words, eliminate all possible sources of ambiguity.
All this is something that ideally should be worked out between the report developers and information users early on in the development process, at the point of business analysis. However, in mitigation it must be said that some ambiguities can creep in later, for example when extra capabilities are added to either reportage or the data environment - or the information user's environment. Be mindful of the ramifications of such changes.
No comments:
Post a Comment