TogetherNow: Improving Search Relevance for Community Benefit
TogetherNow, a non-profit social enterprise, struggled with search relevance issues in their MyWayfinder app's community resource directory. Through their consultation with Gigasearch, TogetherNow significantly enhanced their search relevance and regained user trust.
TogetherNow is a non-profit social enterprise that provides an integrated system of care for its region. It connects a diverse network of service providers to create a comprehensive network of health, education, and human services for its users. Additionally, TogetherNow offers consulting services and implementation support to other communities that wish to implement a Community Information Exchange®. One of TogetherNow's primary goals is to help prevent poverty before it begins.
The following is an interview with Angee Brown, Senior Project Manager, and Jeff Pettenski, Technical Lead at TogetherNow.
Gigasearch helped us prioritize the most critical aspects of our search requirements, teaching us how to craft better queries and adjust the index for more relevant results. - Jeff Pettenski, Technical Lead at TogetherNow
What is your Elasticsearch use case?
TogetherNow's MyWayfinder™ app includes a community resource directory powered by Elasticsearch. Our users utilize the directory to locate relevant services in close proximity to their physical location.
What problems were you facing?
We faced significant search relevance issues due to various factors, which made it difficult for users to quickly find what they were looking for in our app. Given that Rochester is home to many immigrants with preferred languages other than English , we want to eventually support up to ten different languages. We struggled to support all these languages from our Elasticsearch deployment. Additionally, search results needed to be near our users' physical locations, but Elasticsearch's geo search feature did not work well for us out of the box. For example, searches for "food pantry" would not show locations in the correct county.
How did these problems impact your business?
We were receiving negative feedback from our users and beginning to lose their trust. We realized that if we couldn't improve our search functionality, it would be a significant obstacle for our business.
What was most important to you when looking for an elasticsearch consultant?
We sought a-la-carte expertise and didn't want a year-long engagement. From our previous experience with consultancies, there's usually a cookie-cutter template and a lot of unnecessary content before getting to the core issues. We wanted experts who could listen to our problems and apply their knowledge with targeted, tactical solutions.
We spoke to a couple of other consultants before deciding on Gigasearch. The first seemed to be an individual contributor, and their pitch didn't resonate with us. What sold us on Gigasearch was their active listening during the initial call. They showed genuine interest in our problem and domain, and immediately proposed an approach.
How was your experience with Gigasearch?
Before working with Gigasearch, we struggled to implement changes in our system based on the documentation. Elasticsearch is broad and complex; the documentation shows where the levers are but doesn't explain the effects of pulling them in combination. There is significant nuance involved. Working with Gigasearch helped us cut through the documentation and focus on our context-specific problems.
Search relevance is tricky to implement because fixing issues for one search term could create problems for others. We recognized the need for a systematic approach to measure relevance across all critical queries, but our spreadsheet method was inadequate. Gigasearch helped us implement an open-source search optimization framework that allowed us to compare changes against a baseline. This was crucial for us to confidently make changes moving forward.
The entire experience with Gigasearch was helpful and cohesive. Their team was highly accessible - we could easily post questions in a shared Slack channel and receive prompt responses. We appreciated the recorded hands-on sessions and the comprehensive project plan. Their team-based approach exceeded our expectations, with each member offering a unique and valuable perspective.
Closing thoughts?
Gigasearch helped us prioritize the most critical aspects of our search requirements, teaching us how to craft better queries and adjust the index for more relevant results. As we continue to identify and address edge cases, the knowledge we gained from Gigasearch remains instrumental in helping us understand and refine our adjustments.
If you're embarking on a new search project or facing challenges with an existing one, the ROI of working with Gigasearch is exceptionally high. In our case, we reduced our cycle time by at least 75%, and I'm confident that we wouldn't have achieved our objectives without their engagement.
Our search results improved dramatically after working with Gigasearch, allowing us to elevate our search capabilities to the next level.
Community Information Exchange (CIE)® is a registered trademark of 211 San Diego. A Community Information Exchange (CIE)® is a community-led ecosystem comprised of multidisciplinary network partners using a shared language, a resource database, and integrated technology platforms to deliver enhanced community care planning.