Key Driver Analysis

Know which lever actually moved the number

Key driver analysis identifies which dimensions explain a change in a metric. Golden measures each contribution and checks whether the leading relationship still holds after accounting for competing factors.

The problem

Profit's up 22K. The meeting is at 10.

You slice by category, then region, then segment, then category and region together. Two hours later you have a story.

And you're still not sure whether discount drives margin, or whether the discounted stuff all happens to sell through one channel. Get that wrong and a team spends a quarter fixing the thing that wasn't broken.

The solution

Drivers you can defend in the meeting

Golden breaks the change into exact contributions, ranks every dimension by how much it explains, and then stress-tests the winner by holding the runner-up fixed.

You walk into the 10am with the walk from last quarter to this one, the lever that moved it, and a clear note on anything that only looks like a driver.

Step by step

How it works

Pick the change

Start from any chart, a Quick insight, or the What changed board. Pick the two periods you care about.

See the walk

A waterfall stacks each contribution from the starting value to the ending one, colored as growth, decline, or offsetting. The story reads left to right: Binders +9.9K, Copiers +8.3K, Machines -6.3K.

Rank the drivers

Golden ranks every dimension by how much of the variation it explains (eta squared, if you like names) and leads with the strongest one.

Check for confounders

Then it holds the runner-up fixed and measures the leader again. If the effect mostly disappears, Golden says so: "Product Format looks like the top driver of Profit / Sales, but it is largely explained by Sales Channel: holding Sales Channel fixed leaves only 22.0% of its effect."

Split mix from rate

For ratios like margin or conversion, Golden separates a change in mix (selling more of the low-margin stuff) from a change in rate (the same stuff earning less). The two add up exactly to the total move.

Key capabilities

What you get

01
Driver headlines written straight from the numbers, with no AI in the loop
02
Drivers that recompute live as you filter the dashboard
03
Every member labeled: driving growth, driving decline, offsetting, or neutral
04
Plain framing: a ranking describes your data, and Golden says when it can't prove a cause

Why it matters

The wrong driver is expensive

A confident answer pointing at the wrong lever sends people to work on the wrong thing. Golden's job is to name the right one and show you exactly why it thinks so.

Next step