Profit-first growth, without the dashboard theatre.

Retail media grades its own homework harder than Meta does — the platform owns the search box, the shelf and the checkout. How q-commerce ROAS inflates, and the margin-adjusted number to steer by.

Every audit finds the same leaks — zero-order keywords, dead-hour spend, branded over-bidding. The useful question is which leak clears your margin bar for fixing first. A triage method with kill rules.

The platforms differ less on CPC than on take rate, geography and category fit — which means each one has a different break-even ROAS for the same SKU. The allocation math, minus the folklore percentages.

One channel converts in minutes and hands you nothing afterwards. The other costs more per order and builds an asset. The comparison most brands never write down — worked in rupees, including the second order.

Most launch guides tell you what to do in week one. Almost none tell you what result forces you to stop. The measurement-first launch: margin gates before day zero, and scale/hold/kill thresholds you set in advance.

Platform ROAS is the number everyone quotes and the worst one to steer by. Here's how the three levels of ad-profitability math fit together — with the break-even formula and a worked example in rupees.

Cost-per-purchase in Ads Manager is not your CAC. RTO, COD, returning buyers and GST all sit between the dashboard and the truth. Here's the honest formula with a worked example.

Not forty charts — ten numbers, each with a target and a question attached. The Monday dashboard that catches problems while they're still cheap to fix.

Nobody — they're answering different questions. A field guide to the attribution gap: where it comes from, how to measure yours, and the UTM + server-side setup that closes it.

Agents that watch metrics, diagnose funnels and draft budget moves are real. Agents that autonomously run your ad account are mostly a liability. A field guide to the line between.

Shopify, GA4, Meta and your BI tool will give you four different revenues for the same month — every time. The cause is definitional, not technical. The fix is a semantic layer.

Most D2C brands lose money on the first order — that's fine, if you know exactly when the customer pays you back. The cohort math, the formulas, and the point where retention says 'stop acquiring'.

Attribution assigns credit. Incrementality measures causation — what happens when the ads stop. The test designs that work at D2C scale, the iROAS math, and when testing is overkill.

MMM is having a revival — privacy killed the pixel, so statistics are back. But below a certain scale, a mix model mostly redecorates your assumptions. An honest sizing guide.

The browser pixel now misses 20–40% of your conversions — in India, payment-gateway redirects make it worse. What the Conversions API actually is, the setup options, and how to verify it's working.

Blended averages are where truths go to hide. Three cohort views — by month, by first product, by channel — that show whether your brand is compounding or quietly decaying.