graphic for The 2019 Index of Intelligent Technology in HR Tech

 

 
2018-08-28-hrexaminer-photo-img-cc0-via-pexels-ai-blur-codes-report-data-577585-crop-544x329px.jpg
As a researcher I often run into data that tells me that I was simply wrong in my original opinions. And that’s ok, my opinion is just that – a perception I’ve built based on the data that I had available at the time. New data should allow me to re-think my position and I should be fine with that. But actually, I’m not. It erk’s me more than you can imagine to look at data that just isn’t saying what I want it to say – and sometimes if feels  almost like the data is mocking my attempts to make the world a better place.

As we continue into the era of data analytics, big data, AI, and personal tracking devices every HR professional should be prepared to face data that contradicts their beliefs sooner or later. So what do you do when it happens?

Before you rush to throw yourself on your own sword and admit that you’ve been terribly wrong, take a few important steps:

1.        Make sure you are asking the right questions

I once had a senior HR leader for one of the world’s largest banks tell me that they were absolutely   frustrated with their Employee Engagement survey because it kept telling them that they had great employee engagement – meanwhile their top performer retention was down and employee relation complaints were on the rise. In other words, they were asking the wrong questions in their employee engagement survey, and needed to make some adjustments.

2.       Make sure your data is clean

Photo of Stacey Harris, HRExaminer Editorial Advisory Board Contributor

Stacey Harris, HRExaminer Editorial Advisory Board Contributor


If your data is being entered at any point by human beings then it is likely you could have some data cleaning requirements.  A large global pharmaceutical organization I previously worked with spent years trying to implement an enterprise wide HR metrics strategy without success. No one could agree on the field definitions, data ownership, or even location of all the data.  Every enterprise wide standard report they ran received push back from the organization because the data was simply not accurate, or in other words clean.

Their approach was not to simply give up because of unclean data – instead they found 40 fields across all the various HRMS, Talent, and Learning platforms that a governance body felt they could begin working on cleaning up as much as possible. Once they started running the reports, they found further opportunities to clean the data and eventually they had a clear set of reports based on those 40 fields, that senior leaders could trust for decision making. Every journey starts with the first step.

3.       Make sure you are using the right comparison data

Long before I played in the realm of data and research, I was a Learning and HR professional just like most of you – spending my days hacking away at political inequities and preaching, yes preaching, about the importance of investing in best practices and new technologies. I had read all the research (or marketing materials depending on the source), stating that technology and best practices would make my work and life simply better. It would solve all of our (my) problems.  My performance reviews usually included the terms tenacious and goal oriented.  And I often achieved my goal – but found that the new practice or tools simply didn’t work the way I thought they would.

As I matured in both my knowledge and understanding of the industry and data, I came to realize that I was often pushing best practices and even technology in organizations that simply weren’t ready for them.  The organizations I was benchmarking against were either bigger, had more resources, had built an infrastructure, had obtained leadership buy-in through solid CORE HR practices, etc. etc.  My benchmarks and frames of reference were off – a deeper dive into the data would have shown me that I was missing key ingredients for success.

4.       Make sure your sample size is large enough?

If the sample size is too small, it may not represent your entire organization, or more importantly provide enough data to see if specific regions or groups of people are behaving differently. Usually a sample size of 20% of any one group or subset that you are analyzing will give you a strong enough data set to make some good judgment calls.

5.       Make sure your data is viewed over time?

Any time that change takes place in an organization it has an impact on people’s behaviors. New programs, technology, or even policy changes require some level of time to see the real impact. If time is of the essence as it is in most organizations, then make sure that you can look at data across different intervals of time – to see if external factors may be having an impact on your data. Could an economic shift have a major impact in sales and therefore retention over riding other factors?  Could a regional environmental issue cause a problem – such as a blizzard in Ohio that has a major impact on employee tardiness?

If after all this you are still facing data that contradicts your view of the world – then don’t follow your first impulse to throw the computer out the window or worse yet push the information in a back corner. Even if it is telling you that the new compensation strategy you worked on for the last year is having adverse effects, or that employees going through career planning are leaving faster than employees who aren’t, or that the technology you just invested 1.2 million dollars in to change your relationship with employees has only achieved a 3% adoption rate.

Instead, take a deep breath and focus on a plan.

  • Gather more data. A single set of data can cause more questions than answers sometimes. Would a series of focus groups or looking at data from another source help clarify the issue. Is there some additional analytics that you can run to see if your data correlates to other factors.
  • Identify Action Plans. Data is just information – the actions we take based on the data have the real impact. An organization could decide to take drastic harmful actions or constructive positive actions – and in many cases it depends which ideas they have in front of them. Quickly pull together positive actions that can be taken to address the data. 
  • Share the data. Decide immediately who needs to be made aware of the data – even if the data is viewed on a shared dashboard, it is good to get in front of the discussions and questions. The worst political mistake is when someone ignores the data.
  • Be willing to change your mind. This could be the most difficult advice I’ve given. It is easy to say, but much harder to do in real life. We are human after all and sometimes it is easier to just become defensive or frustrated – but remember that data isn’t judgmental.

Data is always telling us something, and if we aren’t willing to listen, then we may just want to think about changing our environment and perspective. Maybe your company isn’t ready to change, but another company is? Maybe our vision of how it should be is masking the reality of how it is. Maybe we just need to take smaller steps to make the world a better place.

 

graphic for The 2019 Index of Intelligent Technology in HR


 
Read previous post:
2018-08-27-hrexaminer-article-learning-by-mistake-by-heather-bussing-photo-img-surfer-sunset-australia-via-debora-cardenas-621886-cc0-via-unsplash-sq-200px.jpg
Learn By Mistake

"It’s okay to fall down, screw up, make mistakes, and figure it out by doing it wrong. Transformation is a...

Close