Color-coded place values so anomalies surface before you read them

The observation

Trade-monitoring analysts spend the day scanning long tables for the one figure that's out of place. Before I designed anything, I sat behind them and watched — specifically, I watched their eyes. They weren't reading rows; they were hunting for shape: a number that looked too big, a column that felt off.

What slowed them down was invisible. Every row forced a small act of mental arithmetic to work out an amount's magnitude — Lakhs or Crores? — hundreds of times a shift. The bottleneck wasn't the data. It was making magnitude legible.

The insight

The anomaly is rarely in the digits — it's in the magnitude. A Crore-sized value sitting quietly in a column of Lakhs is the thing you need to catch, but nothing on screen signalled scale without reading the number first.

So the design question flipped: could the eye catch magnitude before the mind reads the number? Give magnitude a visual identity and spotting the outlier becomes pre-attentive — noticed before you've consciously looked.

The system

I encoded place value in colour across the Indian numbering system — Lakh, Ten-Lakh, Crore, Ten-Crore (L / TL / Cr / TCr) — each magnitude its own consistent hue. An amount now carries its scale as colour, not just as digits.

In a healthy column the colours are uniform. An outlier shows up as a hue that doesn't belong — visible from across the table, before a single number is read. The interface runs the first pass of the scan for you.

The outcome

Analysts began catching anomalies at a glance instead of row by row. Time-to-spot dropped by roughly half, and the mental fatigue of constant magnitude-math went with it.

It's the project I point to when someone asks how I actually think. The real design wasn't the colour system — it was sitting behind an analyst long enough to see that the work was happening in their eyes, not on the screen.