In 2018, we decided that Idealist had the capacity to fully pursue a vision that our staff had dreamed about for a decade: We wanted to gather salary data from across the nonprofit sector and build a reporting tool that could show the salary range for any nonprofit job title.

Idealist’s Director of Audience Success, Emily Hashimoto, led the charge on this project. They outlined the initial product requirements and knew we would need the expertise of Idealist’s resident data scientist, Ann-Julie Rhéaume. They brought on JER HR Group as compensation consultants and formed a Compensation Advisory Board of seasoned HR professionals from nonprofits across the United States who could ensure we were building a useful tool.

In making the Nonprofit Salary Explorer, we knew we had to consider all the ways that compensation reports and resources had been developed in the past. Emily’s market research thoroughly evaluated more than 30 existing platforms and reports, giving our team a sense of how survey questions were typically asked, what data was prioritized, and how results were visualized.

The next major undertaking of the project was to categorize as many common nonprofit job titles as possible. We built a system that allowed for canonical standard titles, along with synonyms that represented essentially the same responsibilities under a different name. Each standard title was then given a level between 1 and 9 based on the seniority of the title, as well as an associated job description. Titles were grouped into families and subfamilies based on common themes.

All told, this process required an intense amount of research and standardization, with oversight from our Compensation Advisory Board and compensation consultants.

We now have a robust system of 2,753 unique job titles (synonyms) for 362 standard titles, all sorted into 27 families with 111 subfamilies.

With the foundation of this title system in place, we wrote a first draft of the questions for our salary survey. We decided to limit the scope of the survey to social-impact professionals who had been employed full-time at a nonprofit organization at some point in the 3 years previous to their taking the survey. This meant we were exclusively dealing with annual (or annualized) salaries—matching apples with apples, without trying to merge the unpredictable “oranges” of part-time roles and hourly wages.

Kyle Marie Kaplan, a project manager at Idealist, researched how we might ask some of the more sensitive demographic questions. While traditional survey design practices have standardized methods for collecting demographic information, best practices have quickly shifted in the last 10 years or so. In many cases, the more established governmental and institutional surveys are lagging behind the evolving understanding of intersectional identities. We knew we wanted to collect this nuanced information so that we could better understand any disparities in compensation practices by gender, race, or other dimensions.

In collecting demographic data, we wanted to be as thoughtful as possible. We made sure to design this last screen of the survey in a way that made it clear that these questions were entirely optional. We included options for nonbinary and transgender individuals and made our best efforts to consider the complexities of race, ethnicity, and people whose background is multiracial. For example, we use the combined framing of “Race & Ethnicity” and allow for multiple selections rather than restricting to a single select field.

In order to ensure the privacy and anonymity of our community, we took great pains to anonymize all data transmitted between or stored in our digital systems. Analytics tracking within the Nonprofit Salary Explorer section of our site is triggered without user IDs or other personally identifiable information. Survey submissions in our database are not stored in relation to specific users but are instead given a unique identifier (a string of numbers and letters). A datetime is stored to log whether or not a user has taken the survey. This date and the date of the submission entry are both slightly “scrambled” by a randomized number of days so that the two cannot be connected to each other. In this way, salary submissions are truly anonymized.

Once we had an initial survey, questions remained: Do these survey questions make sense to people? How can we present them in a way that fosters an engaging user experience? What do people expect to see after they complete the survey?

To learn the answers to these questions, we needed to solicit ongoing feedback and iterate on our designs.

Next: Part 3 - Design & User Research →