Manage critical HR data
Article Type: How to … From: Strategic HR Review, Volume 11, Issue 6
Practical advice for HR professionals
HR data continue to be a problem as well as a boon for most organizations. It is not long ago that HR was bereft of data; the data were of patchy quality; the good data were locked away and as a result HR made decisions based largely on opinion, not evidence. Now, many organizations are overloaded with data –they have counts of virtually everything but, in the words of Albert Einstein,“not everything that can be counted counts and not everything that counts can be counted”.
So, how do we ensure that appropriate data are collected and then appropriately used? In this article we present some ideas on how you can enhance the way you manage critical HR data.
1. Data quality
What are “quality” data? Here are six criteria by which you can judge the data you collect.
Comprehensive. Does the data sufficiently articulate what you need to understand? For example, if you are considering promoting someone and performance matters, does the annual performance appraisal rating tell you enough about their output? Does it tell you how they cope with sudden unexpected changes in demands? Does it tell you if they can sustain performance even when under pressure, etc.?
Valid. Does the data truly describe what it purports to? For instance, does data about why people are leaving truly tell you why they decided to leave, or does it merely tell you why they selected the specific job to which they are moving? Or, are the data sufficiently current? Some things, such as education history, do not change much. However, others, such as personal motivation, can change very quickly.
Reliable. Is the data to be trusted? If you collected it again,would you get the same answer? Has the data been adversely affected by circumstances? For example, does survey data truly reflect the views of the entire population, or was it impacted by the way in which the survey was completed?
Differentiating. Is there sufficient variation in the data for it to inform decisions? For instance, does the applicant’s school really make any difference to their performance?
Useful. The very process of collecting data can indicate to those who supply it that you will make use of it. That can raise unrealized expectations and even risk litigation. For example, if a figure comes out low,are you sure that you would do something about it?
Defensible. Make sure that you can defend why you are collecting the data and the above dimensions of quality of the data. For instance, many organizations collect data about “the risk of loss of top talent”. However, individuals need to trust how this data is going to be used and, if challenged, you need to be able to defend the accuracy of it.
2. Data collection
What do you need to consider when collecting data?
Source. Make sure that you seek data from the best possible sources. For example, change from asking the managers to provide data about employees to asking the employees to tell you their aspirations, mobility and who they think could do their job.
Context. Make sure that the context of the data collection does not adversely impact the data itself. For instance, if you ask employees about their view of pay rates just before annual pay reviews, don’t be surprised if they produce harsher assessments than they do if you ask mid-cycle.
Quality checking. Use contemporary technology to enhance the quality of data. For instance, in addition to normal data validation, it is possible to detect aspects of performance appraisal rater-bias at the time that ratings are submitted – real-time feedback can be presented.
3. Data mapping
The real value comes from linking elements of data together. For example, it is useful to know the perceived competencies of each of your leaders and which have the most engaged staff. But, it is substantially more useful to know that the least engaged staff work for managers who display a particular competency profile.
4. Data analysis
Gain some basic statistical skills and knowledge, such as the following:
Not all big numbers are good. If you have lots of “ready now”successors or people with very high levels of competency, you may be sitting on potential employee boredom and then attrition.
There are right and wrong ways to present data: line charts for trend data or data where adjacents matter; scattergrams for dual scale data; pie charts to show distributions; full scale charts to show absolute values; truncated scale charts to highlight variances and so on.
Not all correlations tell us of cause and effect relationships. Correlations can often arise due to a third variable, and can actually be the reverse of what is suspected or even be merely coincidental.
Appreciate that samples are not always representative. If 80 percent of employees complete a survey, the results of certain questions may be unreliable because all of those who chose not to complete the survey would likely have responded differently to those questions.
You do not need to be a trained statistician to be able to undertake intelligent analysis of data. However, without basic training, it is extremely easy to misunderstand it.
Clinton Wingrove EVP at Pilat HR Solutions.
About the author
Clinton Wingrove is EVP and Principal Consultant at Pilat HR Solutions. He has over 30 years of international HR management experience in-house and in consultancy and is a frequent speaker on the international conference circuit. His specialization is in driving demonstrable and sustainable performance improvement and/or growth to realize true potential – individual, team and organization. He is a member of the CIPD, Chartered Management Institute, IHRM,SHRM and ASTD. Clinton Wingrove can be contacted at: cwingrove@pilat.com
