systematicHR

The intersection between HR strategy and HR technology

,

Doesn’t Matter How Cool HR Technology Is

systematicHR Avatar

You can automate everything, have wickedly cool workflow ((I think I just let everyone know what generation I’m from)), and perfection in the user interface. It’s all meaningless if your data is crap.

1) If data in your core HRMS is crap, then everything else is too. Let’s think about this and what the effects are downstream. A simple mis-entry of an employee as full time rather than part time means that (at least for the short term) the employee was able to enroll in the wrong set of benefits and perhaps got overpaid for the first pay period. While this is all easily fixable, you are now talking about manual fixes and intervention.

A simple data entry error takes on significantly more time to fix than to simply do right the first time around.

2) Perhaps more seriously, if data in your core HRMS is crap, then your metrics are too. Consider that you need to pull metrics about XYZ data, but that data field is not table driven. The 10 data entry clerks you have over time keyed “California” as “CA” Calif” “Cal” and so on. They have also keyed “Canada” as “CA” and “CAN.” You can begin to see the problem.

Unfortunately these are not really errors, but a lack of focus on standards. Avoiding these types of data errors is actually much easier, but the consequences of not cleaning them up are more serious as it impacts what you can use for reporting.

At systematicHR we talk about technology and what we can do with it all too often without addressing some of the core problems that plague us in our daily work. While it’s nice to be a visionary and pretend everything is beautiful, every now and then I’ll come down to earth and remind myself that there are still fundamental problems to be fixed.

Tagged in :

systematicHR Avatar

6 responses to “Doesn’t Matter How Cool HR Technology Is”

  1. Naomi Bloom Avatar
    Naomi Bloom

    In all too many cases, perhaps most cases, the data design of the underlying HRMS is itself so flawed that even high quality data entry processes won’t save you. For example, if duration of employment (i.e. ongoing, fixed term, when actually employed, etc.), nature of work schedule (i.e. full time, part time, on call, etc.), and employment status (i.e. probational, regular, on family leave, etc.) are crammed into an employee status code attribute whose convoluted valid values and related meanings are understood by nobody, not only is data analysis a mess but you’re also faced with the need for mass changes whenever any aspect of these quite separate concepts changes, self service users cannot be expected to understand what all of this means, and the opportunity for errors in automated processes driven by these codes makes regression testing a full time job. Until current HRMS’ are replaced with ones whose data designs weren’t conceived when I was very young and ones which are based on proper as well as current models of the HRM domain, we’re building skycrapers on sand when we attempt to use yesterday’s payroll-centric data designs as the foundation of today’s HRM delivery systems.

  2. Barnaby Fountain Avatar
    Barnaby Fountain

    Your statement, “If data in your core HRMS is crap, then everything else is too.” is right on target. Having experienced the pain of launching a performance management application before our data standards and data cleanup efforts were completed, I can’t stress this enough. It is also very easy to seriously underestimate the amount of effort required to implement HRMS data standards in a very large organization. I was amazed to find how much variation there was in the most basic data elements, like a person’s primary manager. In a multinational organization with multiple HRMS instances, there were many different concepts of organizational hierarchy in use. Some based on managing performance, but others based purely on financial reporting. Implementing standard definitions of each data element becomes a tremendous change management effort. The first step is to prioritize the importance of each data element so the standardization effort can be broken down into manageable chunks. Unfortunately, the project is not “wickedly cool” but absolutely critical to the success of the projects that are cool.

  3. C.M. Peters Avatar

    You touched on a very basic, yet important aspect of HRMS/HRIS and I thank you for that Dubs. All too often it\’s somewhat of an issue in helping my HRIS people understand the importance of nomenclature and simple data entry to accomodate such standardization. But when it comes down to it, no report, no matter how well it is designed isn\’t going to work without proper query tags and valid data.

  4. Dubs Avatar

    We spend so much time talking about strategy or what wonderful new features in HR technology are available, but too often the only time we worry about data quality is during implementations and conversions. Bad data quality can certainly be a spoiler for any strategy. Too often executives will want (as in Barnaby’s example) to launch a new system before all systems are ready.

    There’s a point at which you say “good enough” but I think this has to be a conscious decision that you’ve reached some goal of 95% accuracy or some other metric.

    Unfortunately, as technology professionals, I too often see that when systems go live before data is ready, the HRIS associates are still the ones left holding the clean-up bag and the “we need reports anyway – do them manually” mandates. The strategists still don’t see the pain. Perhaps some of them are reading this?

  5. Jason M. Lemkin Avatar

    This is a great piece. One way to bridge the gap (and what we think about a lot with EchoSign) is to start with as much semi-structured data that is automatically captured as possible. I.e., have the app automatically collect as much information as possible itself (even if it has to be “Canada”, “CA” and “CAN”) and then leverage the manual inputting time & data entry clerks into cleaning up the semi-structured data or adding more identifying data on top of the semi-structured data. This dramatically increases efficiency, and provides a nice back-up plan in case additional data is never inputted at all – at least you have the core dataset automatically inputed by the system.

  6. […] Doesn’t Matter How Cool HR Technology Is HR/HCM Radar You can automate everything, have wickedly cool workflow1, and perfection in the user interface. It’s all meaningless if your data is crap. 1) If data in your core HRMS is crap, then everything else is too. Let’s think about this and …  Read more… You can automate everything, have wickedly cool workflow1, and perfection in the user interface. It’s all meaningless if your data is crap. 1) If data in your core HRMS is crap, then everything else is too. Let’s think about this and … By systematicHR Tags: hr  talent management  hr technology  E-mail |  PDF |  Save |  Blog this! |  Related stories:      |  Bookmark:    |  Follow:          09/07/2006 05:00:10 AM […]