Year 3 Plan

This plan informed SERC's Compass work between the spring of 2023 and 2024. Our active work is informed by our current plan.

Improve outcomes for visitors using the search engine

• Make standardized search interface metrics visible to site managers for all search interfaces. This will include information like that helps individual projects (and SERC overall) understand how people are approaching the interfaces (terms used, facets selected) and what their success looks like (results clicked). Summer.
• Make key search interface behavior changes prioritized by observed use patterns: e.g. implement spell checking, provide better guidance with 0 return results.  Summer. 
• Tune the internal search engine parameters using a standard expert judgement protocol.  This will involve broad data collection from the community on the quality of potential search returns for a range of typical queries. This data set will then be used by a genetic algorithm to find optimized parameters.  The data collection is also an opportunity for community engagement through which we can build awareness of the breadth of the collections. Summer/Fall.  
• Continue to monitor advances in Large Language Model driven search.  Explore low-investment, high potential payoff experiments in using these tools to solve keyword/content misalignment and complex query intent challenges. This area is rapidly changing and deep investment in 'current' tooling is unlikely to be a good long-term choice.  Ongoing. 

Other Cross-Project Navigation and Use

• Investigate the use of and interaction between the SERC front page and the Teach the Earth portal.  Make design changes to improve how these pages guide users to relevant resources. Fall/Winter
• Refine our model of supporting discovery for users interested in broad topics (e.g. plate tectonics) where many resources are relevant.  Explore current use of Site Guides and Teach the Earth 'pop-up' solutions and their role in the Best Bets system for search results. Develop a long term plan for maintaining and updating a core set of these discovery tools. Fall/Winter
• Make refinements to the visual look-and-feel of common site elements to: reduce the sense of dissonance for users moving cross-project, and improve the usability of widely deployed site features that impact discovery.  This may include removal of low-value features that clutter the interface. Fall/Winter. 
• Make site-wide speed/performance improvements to improve ranking within Google search returns.  The 'direct from Google' traffic is foundational in our ability to reach our audience.   Summer/Fall
• Under consideration: (if time allows)
  • Clear labels that directs visitors landing on out-of-date content (old projects as already identified in our earlier inventory work) to more relevant content elsewhere on the site.
  • Explore better visibility (and possible better implementation) for the user ability to mark content as a 'Favorite' and to share a personal list of favorites within the site.  Consider the value of these favorites as a data source for evaluation and perhaps future personalized recommendations.

Partnerships

• Settle on partnership criteria, and pursue a few additional partners in line with criteria.  Summer/Fall
• Explore building a partnership collection with open-access articles from the Journal of Geoscience Education. Summer/Fall
• Explore other partners who may be interested having an EOS-style custom subset of Teach the Earth resources based on their own criteria/resource.  Ongoing. 

Highlight resources of high community priority but low visibility.

• Work with the Teach the Earth committee and other source that reflect forward-looking Earth education community priorities to identify high-priority areas where we have relevant resources.  Summer/Fall
• Rotate identified content into the site-wide recommender and measure impact.Fall/Winter
• Identify key resource types not currently captured by our resource type vocabulary, develop new terms and tag the resources.  This bring better visibility to the resources within search returns.  Fall. 

Evaluation/User Research

• Usability walkthroughs mid-point to compare to book-ends- this can include testing new tools for formative feedback and tuning. In particular, the SERC/TTE front page and activity sheet top page redesign will benefit from testing.  Maybe several iterations to capture as we have new affordances available
• Better characterize the types of DEI resources visitors are looking for, as reported on the happiness survey. Initial data collection in person at EER to inform the design of a short interview protocol. Findings will help leverage the new DEIJ tag in a way that is useful to the community.  
• Refining analytics to understand project success, including a gap analysis of what we are not learning from analytics and how best to fill.
• Characterizing/understanding degree of connectivity of isolated resources (department curriculum matrix, AdvanceGeo pages) and ensure analytics or other metrics are able to demonstrate change/improvements.
• Develop survey to support search parameter tuning 

Community Engagement

• Earth educators participating in our search return judgement collection.  Summer/Fall
• Present Compass discovery strategies at the Earth Educator's Rendezvous.  Summer
• Engage community members in usability testing. 
• Engage individual CDA members around areas of particular expertise.  Fall/Winter