13 notes from the margin
Plain-English thinking on profit, demand, attribution and getting analytics to actually drive a decision. Written by the team that builds the platform. All 13 free to read.
Revenue is not the result. Margin is.
A 4x ROAS and a CFO's loss are both true; contribution margin is the number that reconciles them, and the breakeven-ROAS formula shows why.
Read the 7-minute essay →The three CACs every operator confuses
Blended, new-customer and marginal CAC answer three different questions; the one that governs every scale decision is the one almost nobody computes.
Read the 6-minute essay → MarginWhy your CAC number is probably wrong
Two operators can both quote their CAC truthfully while one number is three times the other; the gap is definition, not error.
Read the 7-minute essay → DemandWeather touches a third of the economy. Here is how much it moves your demand.
About a third of GDP sits in weather-sensitive sectors (NOAA, US estimate); which verticals swing most, and how to position spend 30 to 45 days before peak.
Read the 6-minute essay → DemandRoofing spikes the same week. HVAC lags a quarter.
A single storm moves roofing and HVAC on two different clocks; read the lag and route budget to where demand is heading.
Read the 7-minute essay → DemandThe 30 to 45 day window: forecasting demand before it lands
Weather is predictable for a week and demand for months; the gap is exactly where the campaign budget gets committed.
Read the 9-minute essay → AttributionWhy three platforms all claimed the same sale
One buyer, one order, three platforms each reporting plus-one: the mechanism behind dashboards that claim more sales than a business made.
Read the 6-minute essay → AttributionROAS lies. Profit-led measurement is what is left.
ROAS double-counts sales it never caused and ignores margin entirely; here is the incrementality and break-even math that replaces it.
Read the 7-minute essay → Retention81% of your customers never come back. The second order is everything.
Across 156,000 real DTC customers fewer than one in five place a second order, and the 30-day window decides who returns.
Read the 7-minute essay → RetentionRetention curves: declining, flattening, smiling
A single repeat-rate number hides the only thing that matters; the shape of the cohort curve tells you whether your business compounds or leaks.
Read the 6-minute essay → RetentionEmail and SMS are not retention. They are acquisition you already paid for.
Owned channels are filed under retention and judged on opens, but the data makes them the highest-return acquisition you run.
Read the 7-minute essay → PipelineA $10 lead that never closes costs more than a $200 lead that does
Lead price is what you paid; lead value is LTV times close rate, and pricing leads on value is what separates a profitable pipeline from a busy one.
Read the 6-minute essay → PipelineMost teams never track win-rate by source
A blended win rate hides the fact that some lead sources convert far better than others, and the cheapest lead is often the most expensive customer.
Read the 6-minute essay →More on the way. Guides, benchmark studies and deeper playbooks are in the works, so check back.