Pigeon Insights®

That 35x your email dashboard reports is borrowed last-click credit. Here's how to run a user-level holdout, sized to your consideration period, for the real number.

Most attribution advice describes the nine-figure setup. Here's the measurement ladder by revenue stage, and the cheap proxies to use before the big tools.

A free gift can protect more margin than a discount and lift subscription too. Here are the rules we give clients: pick by spread, never gift the hero, and use mystery gifts to clear overstock.

Most Shopify brands flat-budget every day and leave 15-20% on the table. Here's how we cross-check Meta, Google and Shopify by day, time, age and gender, then reallocate.

What UGC really costs in 2026: Billo, Insense and a creator on retainer, the price-jacking trap, and the cost-per-winner number that actually matters.

The testing campaign that burns budget in three days, and the pack system I use instead: how campaign vs ad-set learnings let new creative climb the CBO ladder.

Operators tipped AppLovin to grab 5-10% of DTC budgets. Here is the real spend floor, who is big enough to test it, and why most brands should max Meta first.

Most agencies explain a bad month with 'you ran out of stock' or 'iOS changed'. Here's the four-part accountability standard to hold any Meta agency to.

AI ad tools don't fail by looking fake. They fail by inventing a stat, a review, or an expert that looks real. The four places fabrications slip into creative.

I rebuilt a 4%-converting product page to follow best practices and watched it tank. Why emotion does the selling, logic only justifies, and how we test both.

ROAS is an output, not a lever. Decompose it into CPM, CTR, CVR and AOV, and you can diagnose any Meta performance drop in about 60 seconds. Here's the tree.

Year-long LTV goals are impossible to act on. Here's the 30/60/90-day cohort target system I use to set actionable growth targets and fund a higher CAC.

Meta value rules let you bid up high-LTV segments and bid down spend-hogging geos. Here's how to set them up and prove the change actually lifted blended ROAS.