We were weighing whether to buy or build a new capability when someone asked a simple question: what do we already have?

Nobody could easily answer this question. Not because people weren’t smart or hadn’t been paying attention, but because the answer was scattered across systems, teams and units that had never needed to line up—until now. We had a decent image of the gap we needed to fill, but the house we were standing in needed to be cleaned up.

That moment changed the direction of the project, but it also changed how I think about what discovery means in a large organization.

The question behind the question

The work began as a strategic exercise: look at what the business was missing and decide how to close the gap. Discovery wasn’t part of the original plan. The need seemed obvious, so the job was to find a solution.

But the more we talked through the gap, the more the conversation kept circling back to data. What did we already collect? Who owned it? Could anyone use it? The answers were uncomfortable. Our own information was fragmented enough that, whatever we bought or built, we wouldn’t be able to make full use of it.

So, instead of asking what capability we were missing, we asked what would actually help the people who’d use it every day, without getting in their way. Their answers showed us which outcomes mattered, and those informed which data needed to be accessible and connected behind the scenes. We started from the person and worked backward to what the organization needed to align.

A small patch of a very large puzzle

I feel like this may be controversial, but I assemble puzzles color first—not the edges—which my family considers roughly as acceptable as pouring the milk in before the cereal. But I’ve never been convinced that the frame is where a puzzle starts making sense. Sort by what pieces have in common, and the picture begins to appear.

Large organizations tend to be edge-first. The frame is the org chart: tidy, bounded, familiar. Each unit assembles its own patch of sky or corner of the barn, and does it really well. We we working on one of those patches, the problems seem clear and the stakeholders take care in solving them.

But the patches do have edges. Other parts of the business wrestle with their own versions of the same questions, collecting their own versions of the same data, and talking to customers who were often the same people we were.

I don’t think this is a flaw in any one organization—it’s what happens at scale. Silos exist for good reasons: they create accountability, focus and clear financial pictures. Yet, learning gets filtered through reporting lines, tools and priorities, and by the time it reaches anyone outside the unit that produced it, it’s become thin or absent.

A few patterns repeat themselves in this space:

• Findings are written for one audience. Research done for a particular team answers that particular team’s questions in that team’s context and vocabulary. Others need to translate it, and most won’t.

• The same need goes by different names. Two units can describe the same customer problem in completely different terms, and both be right. Without a shared language, they won’t realize their pieces fit together.

• The data is siloed the same way the organization is. Even when someone wants to connect the dots, the evidence behind the insight is scattered.

• Nobody owns the question “who else should know this?” It’s no one’s job, so it doesn’t happen.

These patterns are not particularly noticeable or dramatic, which is what makes them so costly. Nobody decides or wants to bury a good idea; it just never travels past the edge of the patch that produced it.

What helped

There was no single fix, but one habit made the biggest difference: sort by color, not by edge.

When we anchored on what would help the people doing the work, the question stopped being “what can we build with what we have?” and became “what has to be true for this to work?” That second question naturally crosses boundaries, because a customer’s outcome doesn’t stop where your unit does. It pulls in data owners, other teams and other priorities, the pieces that share a color even though they sit in different parts of the frame.

The second habit was simpler: translate. Restating what we learned in terms another unit could use, instead of assuming they’d find their way to our reports, turned one patch of knowledge into something that could connect to the next.

The ideas that stay buried

What I keep coming back to is how much value sits in these gaps. The opportunities we found were real, but they lived only in our patch, and I’d be surprised if we were the only ones. The ideas that could change how a business works often aren’t missing; they’re buried somewhere with no way to learn that someone else needs them.

For large organizations, that may be one of the biggest barriers to progress. It isn’t a shortage of talent or insight. It’s a shortage of circulation: the movement of ideas, opportunities and solutions between the places that produce them and the places that could leverage them.

You can both buy and build capability. What you can’t do is skip the work of knowing your customer, or knowing your own house well enough to act on what you learn. Everyone in a large organization is holding some of the pieces. The advantage goes to whoever starts sorting by color.