Pan Macmillan

We partnered with their team to shift their primary websites to a Jamstack architecture. The brief was to  improve work-flow for editors, make content accessible to multiple channels, provide a microservices architecture to allow for a more agile approach to releases, and improve site speed significantly.

The site uses Gatsby, a static site generator framework, whilst editorial content is pulled from Kentico Kontent with external content sourced from various 3rd party services (Good Reads, Twitter, Facebook, YouTube, Instagram). We parse book and author data from an ONIX feed and extracts of books from Pan Macmillan’s own API.

This scale of the sites presented a challenge, the largest site has over 14,000 pages rendered on each build, the majority of which draw content from multiple sources. While there were many smaller sites using Gatsby there were few, if any, sites on the same scale as this build. 

We created custom plug-ins to deal with the data import and implemented scheduled tasks and a caching strategy to aid performance. We processed the book and author file directly which was more performant than querying the existing API and removed a dependency from their process, this allowed for a more agile approach.

* Based on a 6 month comparison post-launch.

A headless CMS means that the content is no longer tied to a single presentation layer and we needed to consider how that content would be used on the existing sites but also how it may be used elsewhere in the future. We worked with Pan Macmillan’s team to establish flexible content models which make the most efficient use of editors' time and avoid duplication of content for multiple streams.

A common disadvantage of static sites is the inability to preview content without rebuilding the entire site. At the time of development, off-the-shelf solutions were not available so we developed a client-side solution for key pages allowing the editorial team to preview content in real-time without having to fully rebuild the site each time.