Statistical Model

Stochastic Frontier Analysis

What is the maximum possible efficiency?

Strengthens Vendor Waste Detector beyond seat-count logic.

The Problem This Model Solves

Benchmarking against average is easy. Benchmarking against perfection is hard. SFA estimates the messy "production frontier"—the maximum output possible for a given input—and measures how far you are from it.

Questions This Model Answers

"How efficient is this channel compared to theoretical max?"

"Are we suffering from inefficiency or bad luck?"

"Which vendors are underperforming their potential?"

How the Model Thinks

It decomposes the error term into two parts: Random Noise (Luck) and Inefficiency (Waste). It draws a curve over the top of the data cloud, not through the middle.

Core Business Use Cases

1Vendor Performance

The Problem

Accepting mediocre agency results.

What It Reveals

The gap between their results and the 'Efficient Frontier'.

Decision Enabled

Renegotiate or switch vendors.

Powered by SpendSignal

How We Use This Model

Powers the Vendor Waste Detector to flag sub-optimal performance that isn't just noise.

Example Output

A scatter plot with a Frontier Curve. Points below the curve are inefficient. The distance is the Efficiency Score.

Works Best When

  • Benchmarking
  • Efficiency audits

Be Cautious When

  • If data is extremely noisy

Stop Guessing. Start Knowing.

See how Stochastic Frontier Analysis changes your budget decisions with a live incrementality audit.

Ask about ROAS, Attribution, or Budget...