Intended audience: DataKind Volunteers
“Great vision without great people is irrelevant.” - James Collins
While the volunteer recruitment process varies across DataKind Chapters, the process always begins by understanding what skills will be required to carry out the project, and documenting these skills and requirements clearly and concisely. DataKind values diversity, equity, and inclusion (DEI) in all our work, and these values must be kept in mind while drafting the volunteer requirements and during the volunteer recruitment and selection process. To learn more about how we incorporate DEI into our volunteer recruitment process, see this Playbook article.
Create a clear job description
Name specific volunteer roles like “Designer”, “Subject Matter Expert”, or “Project Manager”, and write job descriptions with explicit role responsibilities that provide volunteers with clear expectations for the recruitment process. Take advantage of resources like this worksheet for “figuring out the role” and creating an inclusive job description, as well as this template for an applicant rubric. Be sure to use gender neutral language in the job description, replacing his/her with their/they/them and eliminating words that could have a discouraging effect on a group by using this Gender Decoder to identify potentially problematic words. For example, avoid the use of gender-charged words like “aggressive” or “fearless,” which are typically associated with masculine traits, while words like “warm” and “supportive” are coded feminine.
Welcome multiple ways to demonstrate skills
Think carefully about the knowledge, skills, and experience that is actually needed to do the project successfully and be a great volunteer at DataKind. Focus on the specific skills the volunteer will actually use, and be explicit that there are multiple ways to demonstrate each skill. For example, instead of defaulting to requirements like “masters in data science, statistics, computer science, or similar,” when you really just need someone able to identify the feasibility of different data science projects types with certain data, explicitly state the skill needed and give applicants the opportunity to explain how they have obtained that skill, which may or may not have been in a formal educational context. The data science job requirements could look something like:
- Someone who knows their way around the data stack - R, Python, MySQL, etc… You can get the data you need wherever it lives, organize it, and explore it.
- Someone who knows machine learning. You could not only decide when to use a decision tree over an SVM, but you’ve probably written a few custom algorithms yourself.
- Someone with deep statistics knowledge. You think critically about bias in all its forms, and you know how to interpret the results of the models you’re using.
- Someone with enough big data experience that you aren’t daunted by massive flows of data. You can code up enough Hadoop to get things down to a manageable size.
Consider non-technical skills as well
Remember projects need more than technical experts to succeed. For example, a user experience designer with contextual perspective can increase the likelihood that an end product is adopted by the intended community. Name cultural understanding, contextual perspective, and design as requirements for some of the volunteer roles on a team, rather than just looking for people with data science skills. These skills are a top priority in volunteer team selection, as both the context and the perspectives of the users and communities impacted are essential to project success. Include the ability to work well in a diverse team as a qualification for selection for all roles, including technical experts. The ability to work with and value people with different backgrounds and perspectives is essential for everyone at DataKind. Within the job description, be explicit about naming that the recruitment will be prioritizing candidates from a diversity of backgrounds. For example: “We aim to build diverse teams and foster an inclusive environment. We encourage women, Black, Indigenous and People of Color (BIPOC), LGBTQ+, and those who identify with groups who are historically underrepresented in tech to apply. Feel free to spread the word about this opportunity to your community!”
Set clear expectations for the time commitment
Specify what time commitment you are expecting from volunteers. Be clear both about the number of hours per week or month and the anticipated end date of the project (keeping in mind that, while dates can and do shift, continually asking project team members to stay on for another month is generally a poor experience). This not only helps you recruit committed volunteers, but it also helps volunteers stay engaged in their commitment, because we set expectations from the onset to ensure project success.
Explain the volunteer selection process
Detail how you are selecting volunteers. For example, teams at DataKind Global are staffed based on a combination of criteria including technical skills, professional and volunteer experience, and time zone alignment. Additionally, when feasible, we recruit volunteers from the geographic region where the project is focused on and with some contextual or sectoral knowledge. Deciding how to staff projects can include deciding if everyone needs to code in the same language, and (if so) what language it is. Be clear about this - if the product needs to be in R to be sustainable at the partner organization, you should only select volunteers who have R skills.
Team Size Considerations
The above considerations might impact your considerations for team size (typically when recruiting for a DataCorps project). There is no absolute rule to determine the optimum size of a DataCorps team, but in general, having teams larger than 8 can lead to coordination complexity for team leads, and reduction in individual motivation and performance - and therefore productivity and happiness! One way to think about the ideal team size is to review the Defined Deliverables and think about the interdependence of tasks and need for team collaboration. For instance, if the project benefits from volunteers working in discrete workstreams, a larger overall team could work out well! The more collaboration needed between the team members and the higher the interdependence of tasks, the more likely it is that a smaller team will perform better. For most DataCorps projects, the ideal team size is between 4 and 7.
Contributer(s): Caitlin Augustin, Mitali Ayyangar, Benjamin Kinsella, Caroline Charrow, Emily Yelverton, Shanna Lee, Manojit Nandi, Nelsa Peña, Daniel Nissani, Mallory Sheff, Martijn Wieriks
Contact us
If you would like to learn more about us, partner with us, or get in touch, email us at community@datakind.org
Subscribe