How to Research and Implement D&I Metrics
Build a Data Driven D&I Strategy with a Robust D&I Research Process
Once you get past diversity and inclusion basics, the next step is all about finding what works in your organization. This means you need high quality diversity and inclusion (D&I) metrics - ones that help with eliminating researcher bias - so you can track progress over time.
The Diversity and Inclusion Research Process
This five step process outlines how to identify the right metrics for your organization, how to develop a benchmark, and how to tweak metrics as you hit milestones.
1. Identify Diversity and Inclusion Mission
Start with a personal statement for what you want to accomplish or what you’ve been tasked with accomplishing.
Use the prompt: “I want to create a workplace that…”
For example: “I want to create a workplace that communicates openly, has a high degree of trust, and ensures that people can do their best work.”
Take note of acute problems, such as high turnover, to keep an eye on the numbers and tie the solutions to those problems into the goals you’re trying to achieve.
Use this prompt for as many goals as you have. You can merge multiple into one prompt or create one sentence per goal, whichever is best-suited to you.
2. Describe Tangible Actions or Outcomes That Show Goal Success
Define your goal based on the action or tangible outcome that shows the goal has been accomplished.
To illustrate with an example:
Communicates openly → defined as employees feeling they know what’s going on at the company and executives are not withholding information from them.
Has a high degree of trust → defined as employees feeling like the CEO is honest during company town halls and employees feeling they can ask their manager any question and get an honest answer.
A workplace where everyone can do their best work → defined as the “logistics” of the office working for all, such as: no physical barriers, chair / desk arrangements as needed, and people being able to use the bathroom without concern.
These examples may be great for one workplace but may not be right for another one quite yet. One of the best ways to get to specific goals and outcomes that work for your office environment is to ask people via diversity and inclusion surveys or interviews.
3. Pick Relevant Metrics and Survey to Get a Benchmark
Picking metrics means identifying something that, as it changes, signals how well or poorly the outcome is coming along. The more outcomes the organization wants looking for, the more metrics you may need.
For a full rundown of metrics that apply to a D&I success, read our guide to D&I success metrics.
Continuing from the example from above, some metrics might be:
Ranked question, measuring the number of employees who rank a 4-5 for the statement: “I believe major company objectives are discussed openly so there are no secrets about what we’re doing as a company”
Trust in the workplace
Ranked question: “I feel the CEO is completely honest during town halls”
Ranked question: “I can ask my manager anything and get an honest response, even if that’s just them saying they don’t know or can’t tell me because of XYZ reason”
A more inclusive workplace
Initiative-focused metric: “% of bathrooms that are gender-affirming or gender neutral”
Initiative-focused metric: “% of physical accommodation requests accommodated”
Initiative-focused metric: “% of entrances, exits, and inter-floor connections that are fully accessible”
Want to create more open conversations about D&I?
Crescendo Helps With That
4. Improve and Add to Metrics Over Time, Diving Deeper into Specific Areas as Needed
To build further on the example above:
Assume over time the CEO does a lot of work to be more up front and open. Soon, everyone believes the CEO is honest during town halls - that’s awesome! Clearly you’re achieving workplace inclusiveness goals as defined by employee needs.
Extending that metric (tied to the goal of trust) requires asking what’s next on the trust barometer for employees. That could mean seeing the company be more active in the community, or it could entail being more transparent about salary and bonus numbers.
Trust could mean a wide variety of things to different employees, so the key is to ask everyone in the company what trust means for them. While it may not be possible to accommodate every ask, listen intently and prioritize, getting feedback along the way from your inclusion committee.
5. Correlate Metrics to Get the Full Picture
Simply tracking one metric isn’t going to tell the whole story, particularly around ranked-method questions.
Correlating metrics, such as “I feel like I can ask my manager any question and get an honest response,” against demographic metrics like gender or against company segments like department or seniority level helps you get the full picture of how inclusive changes really are.
Without correlations and segmentation of data, it’s easy for bias in research to creep up.
Correlations should consider three levels:
The base metric (a ranked statement, for instance)
A demographic metric (such as gender)
A company metric (such as department or seniority)
As an example, you may notice a marked improvement in the question about manager honesty. However, running a correlation may show that senior females feel their manager is honest but but junior males disproportionately view their manager as dishonest. While the first finding should be celebrated, the latter issue still needs to be addressed so the metric of manager honesty is not yet wholly a “success” for workplace inclusiveness.
Pay Attention to Data Stewardship and Privacy
Since you will be asking for sensitive personal data, ensure you have good data stewardship and privacy practices in collecting and analyzing data.
Limit access to data. Ideally, only one person on the D&I side and one data scientist get access to raw information.
Correlations should be done through employee ID or other semi-anonymous marker. That way if data gets out it doesn’t explicitly name names.
Once anonymized, clear names from raw data sources.
Grouped data needs five or more data points in a cohort to ensure further anonymity.
Data is crucial to track success of any project, and diversity and inclusion is no outlier to that rule. Not only does high quality data help you paint a full picture, but following a repeatable and scalable process ensures that the project can evolve as it needs to over time.
Thanks for reading! If you're interested in more resources like this you can sign up to our mailing list, or if you've got advice/experiences that you'd like to share - we'd love to hear from you!
Crescendo is the diversity & inclusion education app for Slack and we’re on a mission to help companies create more inclusive workplaces. You can learn more here!