Grupo Vocento is a leading multimedia communications group in Spain. It accounts for 24% of newspaper circulation in the country, thanks to its 14 newspapers with national, regional and local coverage, including ABC.es, the third-largest national general-interest newspaper. As a result of the merger between Grupo Correo and Prensa Española, the group also owns various audio-visual media, production companies, and internet sites.
How do media groups improve reader engagement to increase loyalty and monetisation? For Vocento's growing IT organisation, this has brought many technical and cultural challenges.
Application monitoring
User loyalty is a key issue. Ariel Ferrandini, software architect, sums it up this way: "It's very easy to make users leave, and very difficult to make them come back. This has been one of the most significant technological challenges we have faced. It is perhaps the most complicated challenge, because in the end it is very easy to enter a website, read four news items and jump to the next one. If the user is provided with the information in no more than a few seconds, they will not come back. These are things that you have to constantly measure because they have a very big impact," he explains.
With multiple platforms running side by side and some 25 million unique users per month across all newspapers and websites, gaining an overview of systems performance is critical for the 13 development teams. However, before 2018, they lacked a single observability platform that would allow them to see, in real time, the health of their applications and user behaviour. "It was clear that we had a weakness in monitoring our applications. We didn't know what was going on with them." says Ivan Pérez Palomino, software architect.
When problems with one of the platforms arose, they were often reported by the users. With a highly reactive approach, the different development teams worked in silos and lacked the processes and tools necessary to provide context to certain errors. This resulted in longer problem-solving times and a negative impact on user experience.
Less troubleshooting, more shipping
The need for greater visibility across applications and the user experience was only one part of a larger digital transformation. The goal was to reduce time-to-market, which meant improving the performance of different platforms, analysing the behaviour of individual users, optimising architecture to scale in times of peak traffic, and changing the way they worked towards a more Agile model. In other words, achieving the goal of having teams spend less time solving problems, and more time developing new features and services.
"We divided the transformation project into four vectors: Agile organisation and culture, design architecture, quality, and DevOps," says Pérez Palomino. "The first thing we did was to create a snapshot of how we were at that moment, and another of where we wanted to be in one year's time."
One of the key decisions was the creation of the architecture team, made up of Pérez Palomino and Ferrandini. "There is always someone who assumes that role within an organisation," says the architect, "but in creating this area, we were able to begin defining guidelines and standards that would allow us to achieve coherence in development. It's not like before, when you saw code from a colleague and said, “I don't even know what he does.” Pérez Palomino adds: "We have worked hard to document, create framework architectures, and develop a community so that everyone can share knowledge and best practices. On the issue of quality, we focused on training the teams, testing their applications, deploying static code analysis and taking a more DevOps approach by creating pipelines."
Monitoring performance KPIs
One of the architecture department's first initiatives was to get teams to start measuring performance KPIs: error rates, response times, and Apdex (user satisfaction index). The team chose to go with New Relic and Ferrandini and Pérez Palomino began to see the traces of the applications. Ferrandini recalls that moment: "We found errors that we didn't even know were happening. I remember going into New Relic and seeing, almost step by step, what was going on. The platform gives a very high level of detail of what's going on internally, within the code."
The next step was to create dashboards, which the teams began to use to visualise the performance of their applications in real time during the development phase and after each release. Having a common source of information represented a fundamental change in the way several teams worked, which were adopting methodologies, such as scrum, simultaneously. In the past, developers were not able to see the impact of the code they were working on. "Since they didn't have access to production, they could only see a log file," explains Ferrandini.
The volume of information that New Relic offered helped change the conversations that Ferrandini and Pérez Palomino had with the teams. "Thanks to New Relic's dashboards, I was able to say to the multimedia team, for example, 'Look, your library is generating these errors,’" says Pérez Palomino. "In three or four days, they reduced the errors to zero. That's a very positive result. The moment they know, they find a solution."
Google and the mobile experience
Significant traffic on Grupo Vocento's platforms is mobile, especially in Accelerated Mobile Pages (AMP) versions. This Google technology allows websites to load faster on smartphones. The problem? Google uses it as a benchmark for indexing or unindexing pages in its search engine. "When you search for any news or any term on a mobile phone, if there is a page with AMP it will be prioritised," explains Ferrandini."Not only is it prioritised, but Google adds it to its carousel, meaning that when you click on one of these news items, what opens is a specific version that loads without leaving Google. And in that version is our advertising." With AMP traffic growing by around 30%, ensuring that pages are error-free so that they can be indexed is of critical importance to Vocento.
"AMP is a technology that in itself means working on performance, although there may be errors at the time of publication that go totally unnoticed," says Pérez Palomino. "Once the content is published, Google indexes it; but if it sees that we have an error, it begins to de-index us, thereby penalising us."
Before implementing New Relic, the only way to notify these errors was through an in-house validator, which generated noise in the notifications, preventing the teams from knowing about actual errors. In addition, notifications were not real time, which meant that if an AMP page was de-indexed, there was a significant loss of traffic.
Things changed with the implementation of New Relic, explains Ferrandini. "We have defined a custom event in New Relic, which now allows us to see errors in real time. We have dashboards where you can see, halfway through, all the bugs and the history. This has basically allowed us to reduce the reaction time to zero."
"New Relic has given us an ability to see information that we didn't have before. We can see, step by step, what is happening within our applications and also analyse it, extract information and—most importantly—make informed decisions very quickly. We have become more proactive—that is, our teams are becoming self-sufficient in error detection. The ability to create dashboards to communicate health information is a very good result," says Pérez Palomino.
This is especially important for Pérez Palomino and Ferrandini when interacting with teams outside of IT. When traffic drops, they now have real data to show that the reasons are often not technical, but instead, for example, related to the quality of the content. "So many business, management, and any non-technical departments are placing their trust in us, which was lost for many years and had been difficult to regain," says Ferrandini. "But today, not only has it been recovered, but the company's strategic software developments are no longer outsourced."
"Another major benefit is that we can now know the real user experience, instead of having to use a tool such as the one we had before, which was a synthetic tool that ran our pages and those of the competition," explains Pérez Palomino. "Now we have all that information on a single platform, and in a controlled context."
New Relic has given us the ability to visualise information we didn't have before. We can see, step by step, what is happening within our applications and also analyse it, extract information and, most importantly, make informed decisions very quickly.
