Skip to content

Published when there’s a verdict worth reporting.

Notes on forward value, measurement discipline, and what the experiments actually show. Infrequent on purpose: if a piece can't carry something you can act on, it doesn't ship.

Get the next verdict when it ships.

New insights, when there’s one worth sending. Email only, unsubscribe anytime.

Start here

Latest

David Gonzalez-Cameron · July 27, 2026 · Strategy · 11 min read

You've been making bets without the odds

Every real ecommerce decision is a wager: money in, an uncertain outcome, no take-backs once the emails go out. Most operators have been placing those bets without the odds, because the odds did not exist in a form anyone could act on. Here is what the read looks like now that they do.

Read the piece
Retention

David Gonzalez-Cameron · July 21, 2026 · 13 min read

Nobody knows when a customer churns. Here's how to put a number on it.

Ecommerce brands can't see the moment a customer leaves, so they guess with arbitrary 30/60/90-day timeframes. There's a better way to make that guess, and at scale the difference is worth six, seven, even eight figures.

Read
Patterns

David Gonzalez-Cameron · July 16, 2026 · 8 min read

Revenue is not a proxy for the health of your business

A record month tells you money came in. It does not tell you what the revenue cost, whether it stayed, or whether it will happen again. From the operator's seat, the two numbers that answer what revenue cannot: contribution margin and Forward CLV.

Read
Patterns

July 8, 2026 · 8 min read

The returns-fraud number can't hold still

Retail's most-cited fraud figure was published two ways in the same year, from bases $205 billion apart, and its own publisher says the years are not comparable. Why returns fraud is unmeasurable from the top, and how to price it customer by customer instead.

Read
Patterns

July 4, 2026 · 4 min read

Returns are a forward-value problem, not a shipping cost

In apparel, a large share of units come back. Most reporting treats that as an operations line. Netted into customer-level value, it redraws who your best customers actually are.

Read
Trial notes

July 1, 2026 · 4 min read

An email that worked, according to every dashboard

In a controlled trial we ran, a reminder email produced healthy opens, clicks, and conversions for eight straight weeks. It also produced nothing. Here is how we know, and why the difference matters.

Read
Trial notes

June 24, 2026 · 4 min read

The metric you optimize decides which strategies survive

Two emails, nearly identical on the engagement dashboard. One built +$5.20 of forward value per customer; the other built nothing. If conversion rate had made the call, the wrong one might have lived.

Read
Methodology

June 12, 2026 · 4 min read

What a confidence interval is doing on a marketing report

A point estimate tells you what one number did once. The interval tells you what you can safely do next. Operators need the second thing, and almost no marketing reporting carries it.

Read
Patterns

June 2, 2026 · 3 min read

Most of your customers will never buy again

In the DTC books we have scored, most customers buy exactly once, and a second purchase multiplies a customer's expected value several times. The leverage is enormous. So is the waste of chasing it blind.

Read

Run your numbers.

The Diagnostic is free. On one read-only Shopify connection, it scores every customer by what they are expected to contribute over the next 12 months, after product costs and expected returns, and shows where that value concentrates and where it leaks.

Read-only access · No deck required · You keep the analysis either way