Matching Search content to user’s mental models
New York Cares creates opportunities and recruits volunteers at a large scale. More than 50,000 people a year search, browse, and sign up for volunteer projects on the organization’s website.
• Analytics showed disproportionately low traffic from mobile devices to the project opportunities pages. This meant users were potentially avoiding the search experience on mobile devices.
•In addition, users on all devices spent little time on the listings summary view, opting instead to click through to the full page project description view. This pattern is not optimal, because users were exposed to fewer choices when searching for volunteer opportunities.
• A combination of design and content revisions was needed. However, content ownership is distributed across multiple departments at the organization, making it a heavy lift.
• Optimizing for mobile is a must-have for a consumer-facing website.
• The matching process between volunteers and projects is central to the organization’s mission.
This project, shaped by UX insights, was assisted by staff and external content writers, a Salesforce administrator, and a developer.
• Qualitative research. Initial insights on the organization’s unique volunteers were then combined with more generalized academic literature on volunteer motivations.
• Tweaked the design for a better mobile experience.
• Supported content audit, revisions, staff training and guideline documentation.]
• Design and comprehension testing of radically different displays of project opportunities.
This project is ongoing and happening in incremental changes.
In order to improve the search experience we needed to understand more about what user’s were looking for. The project also neatly dovetailed with a Booth-Ferris grant to “look at what types of digitally displayed information motivates volunteer participation”.
For the initial round of research I recruited users that were new to the organization. I suspected that their goals would be fresh in their minds and easier to surface, as opposed to the organization’s veteran volunteers.
• Volunteers used highly transactional language, like “giving back” and “providing value”.
• The range of motivations spanned from the more self-oriented tasks like gaining skills to add to a resume, to the more communal-minded, such as being part of a community.
• This range dovetails with Maslow's well-accepted theory of tiers of human motivations.
See full presentation of research findings
2 Literature review
My initial insights were augmented by the wealth of academic resources on the psychology of volunteerism.
I was helped by a well-accepted definition of the six elements of volunteerism: voluntary action, little to no compensation, longevity, planfulness, non obligation, and organizational context. (Clary et al., 1998; Finkelstien, 2009; Penner, 2002)
The element of planfulness is relevant to the search experience. Search involves a journey through planning, from information-seeking through to informed decision-making.
The academic research confirms that volunteers make decisions rationally. They anticipate a positive emotional response, and consider not only the costs and benefits associated with volunteerism but also the extent to which volunteerism fits with their own personal needs.
Research also shows that the persuasive impact of a recruitment message is greater when it matches an individual’s primary motivations (Clary, Snyder, Ridge, Miene, & Haugen, 1994).
3 Incremental changes to content and design
There were short-term tweaks that would make the project display more efficient and conveying information, and more mobile-friendly. This approach is practical, although eventually a new system display is planned.
I tested at-a-glance comprehension of the new format.
This testing validated the changes. It also strongly demonstrated that potential volunteers create an immediate and instinctive mental match between themselves and the opportunities. This moment of intuitiveness presents an opportunity that the organization was currently missing out on.
Example: Original view, desktop and mobile
Example: New content and design tweaks
Example from testing: New taxonomy, long-term goal for presentation of content (not UI)
The organization has a distributed-authorship model for content creation, which means that many people create web-visible content, often without review. Revising thousands of project descriptions would involve significant training for staff.
For the initial revision we used an external content consultant. We targeted revisions to a four month period of planned projects, which meant revising about 3,500 descriptions.
For the next round or revisions, 8 months later, we used an on staff writer. The writer also created documentation and training for staff so that they could create better-informed project descriptions in the future.
One of the trickle-up results of this project is better staff understanding of writing for digital devices and readability practices. This has impacted content across the website and other channels, including social media and emails.
(During the research process it became apparent that managers lacked understanding of what displayed externally on the website. Their main interface is Salesforce, the backbone of the organization’s management of projects.
• Our Salesforce administrator added tags to identify data as internal vs external facing.
• Project managers had named projects in ways that made access easier for them. We created a new input for web titles, to separate the two types of titles.
• Implementing a content management system and review process is an ongoing effort.
7 Long-term planning
In the future the organization can consider alternatives to the current display system, and to the search process overall. As part of this project I designed and tested a few alternatives. These are generative experiments, rather than firm commitments of a direction.
In this example I tested an ‘eyebrow’ display of project traits. These traits are drawn from volunteer language during research interviews. To scale this type of display would require significant backend and frontend system changes.