Migrating millions of articles: custom CMS ➡️ Arc XP

L'Express: France's Pioneering Weekly

Founded in 1953 by Jean-Jacques Servan-Schreiber and Françoise Giroud, L'Express is a premier French weekly news magazine headquartered in Paris. Known for its center-right stance in France's media landscape, L'Express has evolved into more than just a news outlet, offering lifestyle and job supplements through L'Express Styles and Réussir. As France's inaugural American-style news weekly, L'Express carved a niche alongside major publications like L'Obs and Le Point. Throughout its history, L'Express has played a pivotal role in French journalism, from opposing the Algerian War to challenging the political status quo. With a rich legacy of influential journalists and contributors, L'Express continues to be at the forefront of covering significant national and international issues, fostering informed public discourse.

Step 1 : Discovery phase

During this phase, we essentially need to rush to discover the scope in order to correctly size the project, prepare for the build phase, and identify all hidden obstacles on our path toward migration. The phase was organized in a series of workshops with the Express editorial, product, and technical teams. What happened during those workshops?

• Identifying and mapping all existing entities (Articles, Galleries, Collections, Sections, etc.) and their fields to Arc XP Stories types.
  Analyzing the section structure and discovering the associated templates and pages.
• Reviewing each block of each template and page and deciding if it fits an existing standard Arc XP block. If not, we assess how complex it is to build the custom block on Arc XP content (content being pulled from Arc XP, but the block is custom in its logic), or if the block relies on a feature that was not mentioned.
• Analyzing each third-party (3P) API that we need to integrate: subscriber management, paywall, newsletters, print workflow, news wires, etc.
• Defining and overviewing the SEO & Data strategy: new structure of the website, sitemaps, tagging strategy, Customer Data Platform (CDP) integration.
• Finally, confirming the budget estimation and timeline. Given that media companies are operating under tight financial constraints, it was crucial to avoid any surprises.

Step 2 : Iterative migration build

After completing the discovery phase, we assisted the Express team in securing subsidies from the French Government for online press (FSDP) and then commenced the migration process.

The migration is typically divided into two concurrent streams:

Content Migration

Our team implemented a script that automatically synchronizes any articles created or updated on the legacy, homemade CMS, and migrates them to the new system. This approach prevented the need for a delta migration just before the DNS switch during the go-live. Additionally, we had to write complex migration scripts for each entity, dealing with intricate dependencies and thousands of edge cases, which is common for media organizations with archives spanning over 30 years. We accounted for empty articles, unsupported custom embeds, and articles containing plain HTML or JavaScript. Other challenges included migrating HD print images, outdated and unsupported video players, among other tasks. A final surprise was the discovery of 1,000,000 redirects, which had not been cleaned for the past decade, just a few weeks before launch—despite Arc XP supporting a maximum of 1,000 rules.

Front-End Development

The second team focused on the front end, cleaning HTML structures and adapting them to Arc XP themes, creating custom blocks, and building features not supported by Arc XP on low-code platforms (like WeWeb or Xano). We also created pages on the Page Builder, fine-tuned CSS, and conducted extensive testing. Testing is a crucial component of the project; our Quality Assurance engineers have been writing test suites and scripts for automated testing from day one.

Challenges for Express migration to Arc XP

Our team encountered several significant challenges during the migration, which will undoubtedly inform our approach to future projects.

Middleware for Mobile Apps

Initially, it was decided to keep the existing mobile apps unchanged for a year due to financial constraints, with updates and migrations planned for the next fiscal year. Consequently, we needed to create a middleware that simulated the old CMS's API endpoints, making the apps believe they were still interacting with the original system rather than Arc XP. Although this task seemed straightforward on paper, the reality was different due to the substantial differences in content structure between the two systems. This discrepancy led to an excessive amount of time spent on developing the middleware.

SEO Integration Delays

We made the collective oversight of postponing SEO integration until the late stages of the project. This decision came back to haunt us when we were suddenly presented with over 100 pages of SEO requirements just a few weeks before the deadline. For future projects, we now understand the importance of incorporating SEO from day one and maintaining close, iterative collaboration throughout the project, rather than treating it as an afterthought.

Handling of Redirects

As previously mentioned, the discovery of 1,000,000 redirects came as a major surprise during the final preparations for the DNS switch, just a week before the go-live date. This late discovery necessitated urgent and stressful last-minute adjustments, which could have been mitigated by earlier detection and planning.

‍

Summary of the interview of Anna Sillard by The Audiencers

After the relaunch of L'Express, a quite famous magazine, Anna Sillard, the product owner of the project on the Express side, gave an interview to The Audiencers, a magazine about the media industry launched by Poool, the best paywall system we work with. (Express doesn't, by the way).

You can read the interview there, but I've also provided a short summary here.

https://theaudiencers.com/operations/migrating-1-2-million-articles-to-arc-xp-a-look-behind-the-scenes-as-lexpress-changes-cms/

The interview with Anna Sillard, Product Director at L'Express, covers the extensive technological overhaul the French newspaper underwent after separating from the Altice group in 2020. This migration was primarily motivated by the need to replace their obsolete in-house CMS, Ocari, developed back in 2012/13 with Symfony 1 technology which had become outdated and unsupported. Facing no prospects of integrating with Altice's planned CMS, Cléo, L'Express had to urgently find a new solution.

Anna detailed their decision to adopt the American CMS, Arc XP, influenced by similar moves from other French newspapers like Libération and Le Parisien. The transition involved signing with Arc in December 2021, and the project rollout began in January 2022. They collaborated with code.store, a French agency, to facilitate the integration and communication with Arc's U.S.-based team, avoiding direct interactions during the setup phase.

The migration, completed in December 2022, was not just a technical update but also included strategic decisions like switching from their in-house CRM to Selligent and from their ad-server to Prisma's. These changes were part of a broader effort to modernize and streamline their operations across digital and print platforms.

Despite the challenges such as language barriers, time differences, and managing a significant content transfer (1.2 million articles), the transition was successful. The new CMS significantly improved operational efficiency, allowed quicker updates, and reduced the time required to roll out new features.

Additionally, the team tackled SEO optimization, managed a vast number of redirects, and executed a substantial clean-up of their image and content database to comply with rights management and improve their online presence.

Anna praises the outcome of the migration, noting the editorial staff's quick adaptation to Arc XP and minimal need for intensive training. She also highlighted ongoing collaborative efforts with other newspapers to address common issues with the platform.

Overall, Anna’s experience with Arc XP was positive, emphasizing significant operational improvements and expressing a cautious recommendation of the platform to other publishers.