Directories content migration on wellcome.ac.uk

The directories — pages with information about grantholders are very popular on wellcome.ac.uk. They’re visited predominantly by funding seekers, who want to better understand who’s been funded, where they’re based, and the type and scale of projects they’re working on. Funding seekers mainly get to the directories from scheme pages.

Understanding the problem

Our recent user research showed that people’s experience of browsing the directories is not great. All grantholders and projects related to a particular scheme live on one, often very long page, grouped by year or alphabetically. Users can scroll down to previous years, but it’s not possible to quickly find grantholders using keyword search. And it’s not possible to find relevant projects across different schemes.

The research also showed that funding seekers want to look for grantholders from their own institution and country, which is not easy in the current structure.

When the website was redesigned and rebuilt in 2016, descriptions of approximately 5,000 projects on the old website were lifted and shifted into a field in Drupal 7. As well as being inconsistent in structure, this is cumbersome to manage.

Designing a new structure

We needed to find a way of sorting and filtering content by different attributes like funding area, scheme, institution, country and keyword search. To achieve our goal, we decided that each project, institution and grantholder should be a separate entity in the new structure (see diagram below).

Migrating the content has been one of the biggest challenges of this project, mainly because of inconsistencies in the old data structure.

To make the migration process as smooth as possible, we decided to migrate in different phases. Each phase included a batch of directories with a similar data structure. Migrating in phases has several benefits:

• the rollback process is easy, quick and less risky
• data quality can be tested quickly
• content can be published and tested early with real users.

The first step was to migrate ‘grantholder’ and ‘institution’ to a new taxonomy structure so it could be referenced from a directory page. The next step was to migrate directory pages. A directory page includes ‘grantholder’ and ‘institution’ reference fields — we needed to migrate content for these first to match the content string with the taxonomy name, and get an ID so it can be populated to a field.

The diagram below illustrates how the old content structure is transformed into new content pages.

Listings page

The next part of the project is to create a listings page with the ability to filter using different parameters, such as funding area, grantholder organisation and country, and the ability to search with keywords. We are using the Search API module with Apache Solr to help of do this because we are already using this technology for our site search and the performance is better than a database query.

We have recently made our beta version of grants awarded and project pages live https://wellcome.ac.uk/funding/people-and-projects/grants-awarded. At the moment, it contains a selection of projects from our current funding schemes and we are adding more projects on a daily basis.

We are continuously looking to improve our product and will be doing more usability testing in the future. If you have any comments or feedback, let us know.