Nobody writes case studies about the weeks that got thrown away. You read about the outcome, the polished screens, the happy client quote. What you rarely read about is the moment, a few weeks in, when a team looks at its own work and admits the direction is wrong.
This is a story about that moment, and about why it turned out to be the most valuable thing that happened on the project.
The setup
Our client was a nonprofit association that serves financial institutions. For years, they had delivered value to their members the same way: dense, static reports full of benchmarking data. How you compare against your peers, where you stand, what to watch. The reports were rich, but the format was aging, and their members' expectations were changing with it. So the association came to us with a clear ask: help us turn this into a modern, interactive dashboard.
A clear ask is a comfortable place to start. We had a defined product to design, an engaged client, and a kickoff full of momentum. We ran working sessions, mostly digital, to build context. Who uses these reports? What do they look for first? What decisions does the data actually support?
Then we started designing, and we started in the most natural place for a design team: the product.
The confident wrong direction
Our first direction treated the dashboard as a product problem. How do all of these separate reports condense into a single tool? What does the ideal experience look like when everything lives in one place? We sketched, we prototyped, we aimed at the most complete version of the vision. The honest internal mood during those weeks was simple: we got this.
The problem was sitting underneath the screens. The direction we were designing toward assumed a technical foundation that didn't match what the client's team was building on. We were designing for a future version of their platform, not the one that existed. The concept wasn't bad. It was premature, and premature is its own kind of wrong. A couple of weeks in, that gap between our designs and their technical reality became impossible to ignore.
Pulling back
We had a choice that every consultancy eventually faces. Keep going and let the gap surface later, probably during development, probably at the worst possible time. Or stop, tell the client the direction needs to change, and absorb the discomfort now.
We stopped.
Telling a client "the thing we've been showing you needs to change" is uncomfortable, but the alternative is worse, because a design that can't be built is really just a delay with good lighting. What made the conversation work was that we didn't arrive with only bad news. We arrived with a reframe: instead of starting from the future product, start from what the team can do now, and build a visible path from the current state to the ambition. We laid it out plainly. This is what we've done so far, this is how it feeds into what we need to do next, and this is why the sequence has to change. Same destination, honest sequencing. The client took it, and we moved.
Starting from the data instead
The pivot changed the center of gravity of the whole engagement. Instead of asking "what should this tool look like," we started asking "what is this data, actually?"
That question turned out to be the real project. The association's data was complex. Many metrics, layered definitions, years of accumulated structure that lived partly in documents and partly in people's heads. Before anyone could design a dashboard worth using, someone had to build a mental model of the data that could scale, one that would hold up as new metrics and new views were added over time.
So we got into the data with the client, together. Working sessions stopped being about screens and started being about meaning. Why is this metric defined this way? How does it relate to these three others? What does a member actually need to see to understand where they stand? Slower work, less glamorous than designing interfaces, and exactly the work that made the interfaces possible.
What it cost us
This is a transparency piece, so let's be transparent: the pivot wasn't free. It cost time, and it changed how we staffed the project. When the direction shifted, we rebalanced the team, moving one designer's hours to another engagement while the rest of us went deeper on the data work. It wasn't a crisis, and it didn't require a change order or a difficult contract conversation. But it was a real adjustment, the kind clients rarely see and teams always feel.
What it didn't cost us was the relationship. If anything, the opposite happened.
What changed for the client
The client came into the project with a common idea of what design is: arranging rectangles on a screen and making things look good. By the end, they had watched a design team spend weeks inside their data model, asking why things were structured the way they were and pushing to understand the details underneath the experience.
One of their stakeholders later told us the project changed her idea of what a design process even is. She expected visuals. What she got was a team that wanted to understand the underlying details well enough to build a good experience on top of them. That shift in perception matters more to us than any single screen we delivered, because a client who understands that good design starts underneath the interface will keep making better product decisions long after the engagement ends.
What we'd tell other teams
Three things we took from this project.
The most dangerous phase of a project is early confidence. "We got this" felt great and lasted exactly until reality checked in. Momentum and direction are not the same thing, and the feeling of progress is not proof you're pointed the right way.
Check the foundation before you design the house. The earlier a design direction is tested against the client's technical reality, the cheaper the correction. We now push for that conversation before the first concept gets polished, not after.
Say it as soon as you see it. The pivot conversation was uncomfortable for one meeting. Shipping an unbuildable design would have been uncomfortable for months. Clients can accept a change of direction. What damages trust is finding out you saw the problem and kept going anyway.
The project that came out the other side of the pivot was better than the one we originally planned, not despite the restart but because of it. Sometimes the most useful thing a partner can say is: we were wrong, and this is what we do now.
