Recently I gave a presentation to an international group of students. In the Q&A session, there were several questions about diversity. What had I done in my positions as HR Director to increase diversity? At the business school I gave a somewhat vague and politically correct answer, but the honest answer is: not a lot. Of course I was involved in some diversity projects and there was always a KPI such as “The percentage of females in job levels A-C positions”. One of the students (from Sweden) told about the diversity approach a company she had worked in. The basis of the approach was rigorous and granular measurement. An example she mentioned: it turned out the average length of the job interviews of females was substantially lower (15 minutes or so) than the job interviews of male candidates. They looked at the interviews in more detail, and it turned out that especially in male-male interviews the time spend on small talk (where do you come from? What is your favorite team?) was longer. This gave the male candidates a better opportunity to build a relation with their interviewer.
Triggered by the Swedish student I have given the subject some more thought, and I am now more convinced that if you are really serious about diversity (not only gender diversity, of course) it helps if you start measuring rigorously.
Measuring diversity: lead and lag indicators
Measuring only lag indicators (“The percentage of females in job levels A-C positions”, “The percentage of female high potentials”) should be done but it is difficult to derive actions from these measurements. Below some suggested data elements to collect; some will be easy to collect, some will take more effort (but if you are serious, then you are willing to put serious effort into the diversity initiative). My suggestion would be not to delay action, but start to collect data, preferably not one time, but on a regular basis so that you can start to see some trends. I focus on gender diversity, but a translation to other segmentations will not be too difficult (if you are able to classify the segments).
Data basic elements:
- Some kind of employee ID
- Manager ID (boss of the employee)
- First name/ last name
- Date of birth
- Date in service
- Date in position
- Part-time percentage
- Job title
- Job grade
- Business unit
- Job location
- Functional area
- Talent status
- Performance rating
- Potential rating
- Flight risk rating
- Target bonus
With these basic elements you will be able to prepare some interesting overviews, comparing men and women on various data elements. Examples:
- The % of females per job grade
- The % of females per functional area
- Average salary per job level split by gender
- Average length in position per job level split by gender
- The % of female participants in the high potential program
- Average performance rating men/ women per manager
- Flight risk of men compared to women in similar positions
With some creativity you will be able to collect more data.
Some suggested measures:
- % of female recruiters
- % of female applicants per vacancy
- The % of female Supervisory Board members
- The gender of the people who interview candidates
- The benefit preferences of males/ females (are you really taking the preferences of various groups into account?)
- The % of female receptionists
- Hours worked per week for males/ females
- Participation of females in social after-work activities
- Number of minutes spend in informal discussion with CEO of level C-1 split by gender
- % of females on photographs in your staff magazine/ website/ annual report/ (job)-advertisements
- % of females in your company LinkedIn group
- % of male employees that have taken paternity leave
- Average amount spend on training males/females
- Average number of business trips per person males/ females
- Average Klout-score per job level split by gender
A four step approach
When you start measuring diversity data more rigorously, I suggest a four-step approach:
- Start your analysis by using the data elements that you have readily available. By making a comparison between women/men on the various data elements and combinations of the elements, you will be able to prepare some interesting insights to fuel the discussion in the organisation. Make sure you have some statistical capabilities in your team, to avoid drawing the wrong conclusions.
- Dig deeper. Use students or a task-force of high potential to gather and analyse additional data.
- Develop some hypothesis around the question: what is blocking more diversity in our organisation? Use the gathered data to test your hypothesis, if required gather additional data. Example: Female recruiters will hire more female candidates (more than male recruiters are hiring for comparable positions). If the data (and research) prove your point, you will be able to move to 4.
- Determine what actions can be taken based on your data analysis and research. Set targets, not only on the lag indicators, and start following your diversity KPI’s.
By no means I want to suggest improving diversity is easy. My plea in the article is that if you are serious about making your organisation more diverse, it will help to start measuring rigorously.