Businesses Have a Data Problem. We've Started to Think There's a Bigger One.

For years, the conversation around business software has been fairly consistent: collect more data, connect your systems, build better dashboards, track more metrics.

None of that is wrong. It's been one of the biggest improvements in modern business. Most companies generate more information today than they could have imagined a decade ago.

The challenge is that information rarely lives in one place. Sales are recorded in one platform. Marketing performance lives somewhere else. Finance has its own reports. Operations has another system entirely. Before anyone can make a decision, someone usually has to reconcile different versions of the same story.

We've spent a lot of time solving that problem. It's a problem worth solving. What surprised us was discovering that it wasn't the end of the conversation. It was the beginning of another one.

Once the dashboards were in place, something interesting happened. Nobody questioned the numbers anymore. Instead, they questioned the decision.

Revenue had increased. "Should we invest more in marketing?" Returns had climbed. "Is this a product issue or a fulfilment issue?" Conversion had dipped. "Is this something we need to react to, or will it correct itself?"

The dashboard had done exactly what it was designed to do. It explained what had happened. The meeting was trying to answer something else entirely.

That distinction changed the way we think about analytics. For a long time, we believed our job was to build better dashboards. Clearer visualisations. Faster drill-downs. More flexible filters. Those things still matter. But somewhere along the way we realised people weren't asking for more information. They were asking for confidence that they understood the data well enough to make a call.

One moment from a recent project still sticks with us.

A client asked why we had removed a graph they'd originally requested. The graph wasn't wrong. During internal reviews we'd watched people use the dashboard, and almost everyone looked at that visualisation first. It drew attention immediately. The problem was that it almost never influenced the decision they were trying to make. It was interesting. It was accurate. It simply wasn't useful enough to justify the space it occupied.

Removing it felt uncomfortable. We'd spent time building it. It looked good. But leaving it there meant encouraging people to focus on the wrong thing, and that turned out to be the more expensive mistake.

We saw the same pattern on a different client's finance dashboard. Refund rate and support ticket volume had never been shown together. Placed side by side, the connection was obvious in seconds: refund spikes were arriving a day after ticket spikes, not before. No new data was collected. Nothing changed in accuracy. The two numbers were just arranged so the pattern was visible instead of buried across two screens.

We've started noticing this happens more often than we'd expected. The most valuable improvements rarely come from adding another metric. Sometimes they come from removing one. Sometimes they come from placing two seemingly unrelated numbers beside each other. Sometimes they come from showing less, so it's easier to tell what actually deserves attention.

That doesn't make for a very exciting product roadmap. It does make for better decisions.

Around the same time, the language we used to describe our own product stopped feeling right.

For a long time we called Fenxi Analytics a Business Intelligence platform. That description wasn't wrong, it just stopped matching the conversations we were having, which had shifted from reporting toward judgement, trade-offs, priorities, and questions that don't have a single correct answer but still need one.

"Decision Intelligence" is a term the industry has used for a few years now. We didn't invent it. But most of what gets built under that label is still a dashboard with an extra layer of analysis on top. What we're aiming for is narrower than that: fewer things on screen, arranged around the decision someone is actually trying to make, rather than around whatever data happened to get collected.