Webinar Summary: Using Data in EDI: Driving Evidence-Based Inclusion Strategies

Using Data in EDI: Driving Evidence-Based Inclusion Strategies

One of the common issues clients come to us with is around how to identify what EDI actions they should take that will actually be effective and building more inclusive cultures. And one of the first things we ask them is, “What is your data telling you?”. Their replies are illuminating – sometimes they can’t make sense of seemingly contradictory data, sometimes they have a lot of data but they don’t know what to do with it, sometimes they just don’t have the data, and sometimes they are having trouble getting buy-in to even collect appropriate data.

So, to help discuss these issues and the solutions we’ve used that have worked for many of the organisations we work with, Included ran a webinar hosted by our Data Lead, Raafi-Karim Alidina, about effectively using data in EDI work.

Here are 3 key takeaways from the session:

1. You probably have more useful data than you realise

   Our society is creating more than 400 million Terabytes of data every day. And with global internet usage increasing rapidly (by 22% every year), that data is more accessible than it has ever been. And now, due to free open-source statistical analysis tools (like R) combined with easily accessible large language models (like ChatGPT), even just a basic understanding of statistics and coding can allow you to draw incredibly deep insights from the data around us.

   Most organisations collect tons of useful data, including:

   1. Diversity data on protected (and sometimes non-protected) characteristics, usually housed on HR systems
   2. Workplace data, including personnel and remuneration data like hiring and promotions, salaries and bonuses, account allocation and billable hours, etc.
   3. Culture data, such as data on employee engagement, job satisfaction, and even some high-level inclusion data.

The problem is that most of this data is only used for superficial and urgent analysis, such like legal reporting for regulatory bodies, financial analysis for annual reports, and sometimes for marketing and basic performance evaluation. In fact, most data that organisations gather is either obsolete, unused, or under-analysed.  This obsolete and unused and under-analysed data (also called “dark data”) makes up 85% of the data organisations have.

This means there is tremendous opportunity to leverage data that we’re already collecting and need little resource to make use of.

2. Combining data that often sits in different systems or that is accessed by different departments is an opportunity worth taking advantage of

   For many organisations, the different data they gather exist in different places, sometimes on different systems. This means that anyone seeking to do more advanced analysis likely doesn’t have access to all this information and may not even know what data exists. But bringing this data together can help us find deep and incredibly useful insights that can drive more effective action around equity and inclusion.

   For example, with one former financial client, they believed they were being extremely equitable, despite their gender pay gap analysis showing that men were paid significantly more than women in the company.  This is because their bonuses were entirely based on how much revenue they brought into the firm.  If you didn’t get the results, you wouldn’t get the reward. So the firm felt they were being extremely fair and meritocratic.  However, when we analysed their detailed personnel data combined with their HR data, we found that the highest value accounts were mostly being allocated to men, while pro-bono projects and internal company projects like running employee networks were mostly given to women.  As a result, there was less opportunity for women to bring in more revenue than the men, and this was baked into their system.

   By combining data from different departments in different systems, we were able to surface an equity issue the organisation didn’t even know they had.  And this is a common problem across a variety of organsations we work with.  By bringing together different sources of data, we’re able to understand much more deeply what is actually going on in an organisation or team, which provides valuable opportunity to design targeted interventions that drive equitable outcomes.

3. Collect behaviour data to understand inclusion
   

   Most organisations collect some data around inclusion in their organisations. Usually this comes from engagement surveys that ask questions like, “Do you feel like you belong at this company?” or “Do you feel like your voice is heard?”  This can be useful for getting a high-level sense of how people feel in the organisation, but it is limited in that it doesn’t tell you why employees feel that way.  They are also often hard for employees to answer.

   However, by collecting more specific information about the behaviours that make people feel included and excluded, you can gain much deeper insights into not just whether people feel their voice is heard, but what behaviours are making them feel that way.  Moreover, by combining this with demographic or personnel data, you can better understand who is feeling particularly included or excluded.

   Leveraging regression analysis and difference of means analysis (both of which can be done on free open-source software like R), this behavioural data combined with personnel and demographic data can show you which non-inclusive behaviours are happening in your organisaiton, what behaviours are affecting which outcomes you might care about (like retention, promotion rates, belonging, etc.), and who is experiencing those behaviours the most.

   Armed with these insights, you have multiple ways of developing targeted solutions:

   1. You can focus on the behaviours where there is the greatest room for improvement (the low-hanging fruit)
   2. You can focus on the outcomes you are trying to achieve, and design interventions around the behaviours that have the strongest effect on those outcomes
   3. You can focus on the groups that are feeling the least included and develop targeted interventions to provide them support

Regardless of how you might choose to prioritise which of the 3 approaches works best for your organisation, all of them will be driven by your own data from your own people. This means they are much more likely to be successful, effective, and sustainable over the long term.

At the end of the day, this is why leveraging EDI data is so important. We can’t do everything – we don’t have the time, the personnel, or the resources to invest in every potentially useful intervention.  However, if you leverage the data you have, combine different data sources to draw deeper insights, and collect behavioural data to truly understand how your people feel, you will be much better set up to build more equitable, inclusive, and effective workplaces.

~Raafi-Karim Alidina