Every dashboard and analytic I have ever delivered has been 100% accurate in every way shape and form. If you believe that, then you’d also believe that I can bike as fast as Lance Armstrong. All jokes aside, I would like to be able to deliver a 100% accurate dashboard but I accepted early on that this is nearly impossible and that this is not a problem just in HCM but across all areas. Just to clarify, I am not talking about the simple analytics such as ‘performance rating distributions’ or ‘jobs currently not in use’ which can be put in an excel spreadsheet. Those can and should always be accurate. I’m talking instead about complex analytics such as ‘headcount trending’ or ‘workforce profitability’ which require trending, complex calculations, roll ups by multiple hierarchies, etc. Generally, complex analytics are hard to do and simple to get wrong.
In your first iteration of a headcount report, it is unlikely you will be able to deliver the full picture. The mainstays of a 1st iteration of a headcount report are; headcount by supervisory and/or department hierarchy, types of movement in the organization (e.g. hire, fire, transfer, loa), employee vs. contingent, and maybe some trending. This is not by any means a complete picture and to get that, you’ll need to add things like all worker types (e.g. vendors, partners, etc), matrixed organizations, headcount by project, etc. These are things that you may never be able to deliver on.
Just to make you feel a bit better, you will also with 100% certainty, have data integrity issues and most likely a calculation error or two. Getting manager to manager transfers within the same department to roll up the organizational hierarchy correctly to the VP who wants to see Internal Mobility is a surprisingly tricky thing to get right.
With this in mind you will need to head off the inevitable calls telling you that your dashboards are inaccurate and thus unusable. Although it may seem counterproductive, you should tell your consumers that there will be errors but you’re working on corrections and you can do this by/ through:
- a formal statement someplace in your dashboards which informs your consumers on how to report an error
- a link to a document that has all the definitions, calculations and data sources for all of your analytics within the dashboards
- While pretty graphs are great for instantly communicating lots of aggregated and calculated information; for determining root cause analysis however and for ultimately taking corrective action, the consumer needs to look at underlying information (e.g. a simple turnover graph that can drill into a table separating turnover by locality, job, performance ratings etc).
- associating accuracy and usage information to each of your analytics; similar to the score you’d see when deciding what latest mobile app to download for example. This way you can focus your efforts on either correcting an issue, various improvements or discontinuing now irrelevant dashboards and analytics
- placing your release or revision date someplace prominently on your dashboards and linking it to your release schedules
Complex analytics and dashboards are hard and I wish you luck with them - knowing and admitting publically that there will be problems with these, will go a long way in giving you credibility and leeway in getting them corrected.