Recipe

The Margin Bridge Most Tools Can't Build

Margin data does something revenue data doesn't: it crosses zero. Segments flip from profit to loss, percentages explode on small bases, and standard attribution math quietly falls apart. FactorPrism® was built for sign-crossing data.

The Challenge

"Margin Compressed 240 bps. Where?"

Margin movement hides in the interaction of many segments — some profitable, some loss-making, some flipping between quarters. The question isn't just which segments moved; it's where the compression lives: a broad-based cost force, or three specific product lines underwater.

The Problem

Zero Breaks the Usual Math

−

Percent-change math fails at zero A segment going from −10K to +10K has no meaningful growth rate; tools either drop these segments or produce nonsense.

−

Built-in functions rule margin out by definition Snowflake's key-driver function requires a non-negative metric — margin analysis is out of scope before you start.

−

Averaging across mixed signs misleads Profitable and unprofitable segments offset inside any aggregate, so "margin roughly flat" can hide two large opposing stories.

The FactorPrism® Way

Sign-Aware by Design

FactorPrism®'s engine handles negative values and sign-crossing series natively:

Gains separated from declines Offsetting stories stay visible instead of netting out to "flat."

Magnitude and direction, at the location where each factor acts A cost force compressing every product line vs one line that flipped to loss — reported as different things, because they are.

Reconciles exactly to your reported margin change Finance-grade and auditable — the bridge ties out to the unit.

Works for any sign-crossing metric Net cash flow, net adds, P&L lines, contribution margin — same engine, same guarantees.

Margin tables are usually one query away

Connect a view and see your bridge — install from the Snowflake Marketplace.

Get it on Snowflake Marketplace