Senior leaders don’t usually see database queries. What they feel instead is the knock-on effect. Teams waiting. Dashboards hanging. Meetings where someone says “it’ll load in a second” and everyone stares at a spinner. That was the situation here. A data-heavy platform built for managing sensitive heritage records was solid in principle, but everyday use felt like dragging an old car uphill. Developers were blocked. Non-technical users were frustrated. Delivery slowed to a crawl.

Download the full case study here.

The system sat on top of an open source platform used to manage cultural heritage data. Think historic buildings, protected sites, planning records. Loads of structured information, used by archaeologists, planners, and public bodies. Under the hood, it relied on a database layer that translated human-friendly actions into database queries. All normal enough. The trouble came from how that translation worked. Every time someone searched or filtered data, the system pulled back far more than needed. Even simple requests triggered heavy queries. Picture asking for one file and getting the whole filing cabinet dumped on your desk.

Pages took 25 to 30 seconds to load. Filters barely helped. Developers worked around it by loading everything, then sorting it manually. Bit of a pain, that. For leaders, the risk wasn’t just speed. It was confidence. When tools feel unreliable, people stop trusting them.

The root problem lived in the query layer. It was doing too much, all the time. Models were bloated. Logic repeated. Documentation thin. Any change felt risky because teams around the world relied on the same workflows. Replacing the whole platform wasn’t an option. Pausing work wasn’t either. The goal became clear. Make it fast again without breaking anything. Galvia Digital was brought in to focus on that one thing. Not a redesign. Not new features. Just performance.

After digging through how data was fetched and filtered, the bottlenecks were obvious. Fixing them meant being careful, testing constantly, and knowing when to leave things alone.

The solution focused on slimmer data models and smarter filtering. Queries started returning only what was actually needed. Range filters worked properly. Duplicate logic was stripped out. A compatibility layer kept everything else running as before. No retraining. No workflow changes. The difference was immediate. Queries dropped from half a minute to a few hundred milliseconds. Pages loaded instantly. Teams stopped waiting and started moving again. For decision makers, the win wasn’t the numbers. It was momentum.

Developers shipped faster. Users trusted the system again. Future changes felt possible, not risky. Not saying every project goes this smoothly, but this one showed how targeted technical work can unlock real business value when it’s done with restraint.

Download the full case study here.