Balancing quality and stability with the push to deliver new features is never easy – sometimes, you need to hit reset. In this blog post, our CTO Benny Olsson, takes an honest and transparent look at the challenges Norce faced in late 2023 to early 2024, and how we worked to turn things around.
Quality and stability are always top-of-mind for any SaaS company, including Norce, and improving them must be seen as a continuous effort rather than a periodic exercise. However, despite best efforts, you may still find yourself in a situation where almost every change introduces new problems, where each release is riddled with issues affecting customers and partners alike, and where your engineering teams feel like they’re navigating through uncertainty.
If you just nodded in recognition, you probably worked with Norce in late 2023 and early 2024. We won’t sugarcoat it - those weren’t our best days.
In 2024, we took a hard look at how we build, test, and deploy Norce Commerce. While quality has always been a priority, we dedicated six months to a structured, cross-team initiative to improve stability, reliability, and performance across the board. The results? More predictable releases, a stronger feedback loop, and a significant reduction in production issues.
As we worked to roll out our next-generation platform, Norce Commerce, while simultaneously maintaining Storm and Jetshop, we began seeing an increasing number of stability-related issues in both Storm and Norce Commerce.
Since Norce Commerce was originally based on Storm, some of the software-related issues – like lingering bugs and inconsistencies – carried over. However, many of Storm’s biggest challenges weren’t just in the software but in its aging infrastructure, which struggled to keep up as demand grew.
Norce Commerce was built on a modern, more flexible foundation designed to scale smoothly as our customer’s needs evolved. While this solved Storm’s infrastructure limitations, it didn’t automatically fix every inherited software issue – something we had to tackle head-on. However, because the two platforms share a common foundation, many of the improvements we made in Norce Commerce could also be merged back to Storm – meaning our focused effort on quality and stability benefited both platforms.
When analysing the situation, we also identified ways to improve our workflow and our release process. One major change was introducing a dedicated Release Candidate environment, where both internal and external stakeholders could test a release before it was pushed to production.
We also recognised the need to shift from a continuous deployment approach to a more structured, sprint-based deployment model. This allowed us to test the product as a whole before committing to a production release, ensuring greater stability and predictability.
Our developers are highly skilled and experienced, so we knew from the start that the issue wasn’t a lack of technical ability. Instead, the real challenge was that we, as a company, fell short in creating the right conditions for them to deliver quality to the fullest extent of their capability.
With that in mind, our improvements focused on three key areas: culture, tools, and processes – because without addressing these, any attempt at improving quality would fall short.
To make a real impact on quality and stability, we evaluated and refined many aspects of our engineering and product workflows. Three key areas stood out as having the greatest effect:
At the same time, we also launched a parallel effort to support our solution partners in migrating customers from Storm to Norce Commerce. This was an important initiative – not just because Norce Commerce offers a more stable infrastructure, but also because reducing the load on Storm helped improve stability for the customers still using it.
Before the initiative, I’m sad to say we experienced several major incidents every month. In this context, a “major incident” meant an issue that affected multiple customers and disrupted their ability to use our platform.
As we approached Black Week, roughly two-thirds through the initiative, we knew this would be a critical test. Thanks to both the improvements made within Engineering and the customer migration effort, Black Week passed by without any serious interruptions. December followed suit, remaining calm, and since then, we haven’t had any major incidents on the scale of what we regularly faced just a year earlier.
That’s not to say that we haven’t had a single incident – but the results speak for themselves. While there’s still work to do, we’ve already come a long way toward delivering a product that “just works”.
It’s always difficult to distil the most important lessons from an initiative that spans months and fundamentally changes how you work and deliver your product, but I’ve identified a few high-level takeaways that were critical to our success.
As I mentioned to begin with, improvements to quality and stability must be seen as a continuous effort. Looking ahead, we’re finally at a point where we can start rebalancing our priorities – tackling our long list of feature requests while still dedicating time and resources to maintain quality and stability.
I’m incredibly proud of what we’ve accomplished during our six-month focus on quality and stability. I look ahead to 2025 and beyond with great expectations for what we, our solution partners, and our customers will achieve with Norce Commerce.
What are your thoughts on achieving high-quality and stable SaaS products? If you worked with Norce before this initiative, we’d love to hear your perspective on our journey. What’s your take from the outside? Is there something you think we should have done differently over the last six months? We’re curious to hear what you think!