When it was time for Germany’s Kölner Stadt-Anzeiger (KStA) Digitale Medien to reinvent its Web site, it didn’t make incremental changes; it basically started over and reinvented its traditional local site. During this week’s Webinar, INMA members heard from Christine Schönfelder, CPO, head of Digital Competence Center, and Timo Schillinger, senior product manager, about how they orchestrated and implemented what they call their Big Bang relaunch.

Such a drastic approach was essential, they explained, because the tech stack and CMS were nearing end of life: “We had nearly no maintenance ability anymore,” Schönfelder said. “We were running out of updates and licenses, so there was really no way to go forward with this tech stack.”

The limitations were far-reaching and prevented KStA from developing new types of content, improving distribution, implementing AI, integrating personalisation, and much more. After “endless discussions,” the company decided the best approach was also the most challenging and riskiest.

“We put eight projects in one big relaunch,” Schillinger said, noting that each of the eight projects was a sizeable challenge in itself. “It would have been a lot of work to do just one of these things. We tried to put everything in one relaunch and to do it in less than eight months and with a fixed date.”

Schönfelder explained they took such an extreme approach because they needed to grow quickly and increase digital subscriptions as print declined.

“We knew that the time to invest in technology, invest in products, invest in ourselves, is limited,” she said. Once that was done, the company could focus on the market, users, readers, a new business model — and a brighter future. Rather than spread that spending and development out over a few years, KStA decided to go all in: “We put that all together and said, OK, this is what it is going to cost, and after one year, we [will be] done and ... ready for the race in the future.”

Creating the Big Bang theory

To introduce a new tech stack and ramp up the development team, Schönfelder said they began working internally to increase knowledge on the team but soon found disagreement amongst members of the development teams and the approach they should take. The team also said it could achieve the goals of the Big Bang launch — but not if it had to introduce a new tech stack as well.

Rather than change the ambitious deadlines they had set, Schönfelder and Schillinger chose to divide and conquer.

“We decided to build two Web sites at the same time with just a short latency and launch the first one on the hard deadline of one year on the old tech stack,” Schönfelder said. “And while building that, we separated a second team that built exactly the same Web site on the new tech stack with only six weeks between those two launches.”

The first redesigned site, its minimum viable product (MVP) launched in November 2022 on the old tech stack and the second was rolled out at the end of January 2023.

Schillinger said that meant making tough decisions, such as what features to cut in the interest of time. By cutting out its “wow” feature, it could stay on time and, quite likely, reduce mistakes. But that doesn’t mean that everything went smoothly when the first site launched.

“When we started with the first launch, we had a lot of CTR problems on our homepage,” Schillinger said, adding that the CTR dropped from almost 60% to 40% on the homepage. With the clock ticking and the launch of the second site approaching, the team had to resolve the problem. It also continued doing user research, asking for feedback from readers about what was working and what they could do better.

The power of two

While the two-site solution brought plenty of headaches, it also held some unexpected benefits, according to Schönfelder: If they had done a completely new site and tech stack at the same time, it would have eliminated a safety net that allowed them to roll back and rely on the old tech stack when needed.

“We had this headache of ‘What if the whole thing is going to explode and we can’t roll back because we completely cut off all the roots that we had before?’,” she said. “Therefore also this decision of taking a two-step scenario — first front-end CMS on the old tech stack and then afterwards the new tech stack on a six-week-old front-end CMS … gave us the opportunity also to do this rollback.”

The third step, which is what teams are currently working on, is to “get in an iteration mode so we can start rebuilding and start improving the things we envisioned,” Schillinger said. He pointed out it would have been impossible to create a product that was everything it needed to be, so they focused on launching it and then improving it. Improvements include an AI recommender on the homepage and hyperlocal personalisation. Their efforts have paid off, with subscription conversion rates subscriptions nearly doubling. The company’s 200 editors have been onboarded to the new CMS and are happier than they were before.

Users are happier, too, enjoying faster page speed and frictionless usage on the homepage. A redesigned starting page offers more reading incentives with a wider range of articles, different kinds of teasers, and easy access to subscriptions and log-in.

Importantly, KStA was able to create its own advertising vertical and take ownership of its programmatic and display advertising for the Web site. It created its own ad server and started working with outside customers, becoming a marketer for other publishers.

“We built a really high-quality and high-end advertising portfolio because of this relaunch,” Schönfelder said.  

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .