Prior & Effect

Prior & Effect

Bayesian Measurement Systems
Where your priors meet reality.
You've been correlating.
You've been attributing.
You've been guessing.
Most measurement systems confuse precision for truth.
Notes from the people figuring out which ones matter.
"All models are wrong, but some are useful."
-- George Box
Bayesian Updating
A prior meets new data and becomes a posterior.
The Approach

Curiosity. Rigor. Clarity. Trust.

01

Curiosity

Essays start with questions we genuinely wanted to understand, not pre-packaged answers.

02

Rigor

Peer-reviewed sources, real math, defensible frameworks. Shown, not hidden.

03

Clarity

Make measurement legible to the practitioner, the VP, and the CFO. If it can't be explained simply, it isn't understood well enough.

04

Trust

Write so the reader can verify everything. No black boxes, no hand-waving.

Adstock Decay
A media impression doesn't act once. It echoes.
The Mission

Radical transparency, so marketing mix modeling can be understood and adopted within your organization.

MMM isn't actually that mysterious. The industry just made it feel that way. We're writing in the open so the whole community of marketers, analysts, and operators can see inside the black box, push back on it, and build better intuitions about their own measurement.

No black boxes. No gatekeeping. Just the work, shown.

The Creators

Rigor meets reality.

Academic depth, operator instincts, and a shared excitement about where the future of measurement is headed.

Contributor
Seth Brand

Eighteen years across nearly every side of marketing and measurement. Quantitative research at Fortune 500 CPGs, global demand generation at a multinational manufacturer, and now growing the measurement consulting practice at Ovative Group.

Connect on LinkedIn →
Saturation
The first dollar is the most powerful. The last dollar is the least.
The Journal

Notes from the posterior.

Field writing on Bayesian measurement, MMM craft, and the commercial math of modern marketing. Written by practitioners, not marketers.

Issue 02 · Measurement Fundamentals · Part II of IV
Prior & Effect

The Prior as a Strategic Asset

Where institutional knowledge enters the model, and why it matters more than most practitioners realize.

Hedi Moussavi, PhD · 15 min read
Prior-Posterior · Well-Calibrated
Posterior moves defensibly
Adstock Decay · α = 0.5
Moderate carryover
Saturation · Flighting Interaction
In-flight vs. annualized
Issue 02 · Measurement Fundamentals · Part I of IV
Prior & Effect

Building a Model Worth Believing

What it takes to build a Bayesian MMM whose outputs can be trusted -- by the data science team, the measurement consultant, and the CMO.

Hedi Moussavi, PhD · 10 min read
Posterior Width · 12 Months
Wide uncertainty
±6.1pp
Posterior Width · 24 Months
Moderate confidence
±3.5pp
Posterior Width · 36 Months
Tightest posterior
±1.9pp
Issue 01 · Measurement Fundamentals
Prior & Effect

Did Your Marketing Actually Work?

The case for incrementality and causal inference -- and why correlation is never enough.

Hedi Moussavi, PhD · 25 min read
ROAS vs. Incrementality
The Fool's Gold trap
Fool's Gold
True Value
Mixed
Credible Interval · Time on Air
Uncertainty narrows with data
Wk 1 Wk 52 ±wide ±tight
Adstock · Geometric Decay
How long the signal lasts
Paid Social
Display
TV
More essays in the series coming soon.