Read the Growth Engine

The Growth Engine answers one question: which channels deserve more money, and which are quietly costing you? It is the screen where acquisition spend meets real profit, so a channel that looks strong in an ad platform can be shown for what it actually returned.

How the screen is organised

The default view is Channels, and it is the one that matters most. The Google, Meta and Microsoft tabs are zoom-ins on a single platform’s campaigns; they do not replace the channel view, they drill into part of it.

Every view on the page uses your Shopify direct-to-consumer orders as the source of truth — not what the ad platforms report. That single decision is why the numbers here can disagree with your ad accounts, and why they are the ones to act on.

The Money View

The channel table is the heart of the page. One row per acquisition channel, with the money on the left and the judgement on the right:

• New customers and ad spend — the raw acquisition inputs.
• 90-day profit : CAC — profit per acquisition dollar. See how it is calculated; note it is a trailing figure that ignores the date selector.
• Revenue returned % — returns, which quietly move channel profitability more than most operators expect.
• MER — revenue per advertising dollar, the fast directional read.
• Role — whether the channel finds customers, closes them, or both.
• Attr. confidence — how well-evidenced that row’s channel assignment is.
• Recommendation — a one-word call: Scale, Maintain, Pause, and similar. Hovering gives the reasoning.

Read across a row rather than sorting by any single column. A channel with a strong ratio and low attribution confidence is a different proposition from one with the same ratio and solid evidence behind it — the Recommendation already folds those together, which is why it is the most useful column on the table.

The Unknown / Blended row

Directly beneath Organic/Direct sits a row for orders MarginOS could not attribute. It is not a channel and there is nothing to spend on it. If it is large, fix tracking before you act on anything else on this page — every comparison above it is being made on a partial picture. Full detail in how MarginOS assigns each order to a channel.

Rows that are shown but not counted

Shopify POS, marketplace and wholesale rows appear for revenue and profit context, greyed and money-columns-only. They never affect direct-to-consumer acquisition cost or the profit-to-CAC ratio, because none of them is a customer you paid to acquire.

The Journey View

Opened from View journeys on the Money View, this shows how customers move across channels before buying — which channels find people and which close them. Its practical use is to stop you cutting a channel that rarely gets the credit but repeatedly starts the journey.

How to use it

1. Check the Unknown / Blended share first. If a large share of orders is unattributed, the rest of the table is provisional.
2. Rank channels against each other before comparing any of them to a target. Relative order is the decision you can act on today.
3. Read the Recommendation with its reasoning, not the ratio alone.
4. Drill into a platform tab once you know which channel to investigate.

Common questions

Why do these numbers disagree with Google or Meta?

Because ad platforms report the revenue they believe they caused, and each claims the same order. Summed, they routinely exceed what your store actually took. The Growth Engine counts each customer once, against the profit your store really recorded.

A channel shows spend but N/A for profit-to-CAC. Why?

The spend has not produced attributable new customers yet, so there is no cohort to measure. That is reported as N/A rather than a zero ratio — the wasted spend is still visible in the channel’s own columns.

Why does the profit-to-CAC column not respond to the date picker?

Its window is part of the metric’s definition rather than a view setting, so it stays a 90-day trailing figure. The full reasoning is here.

Should I act on MER or on profit-to-CAC?

Profit-to-CAC. MER is revenue-based and cannot see product costs, so it tells you something changed rather than whether it was worth it.

What do I do with a Pause recommendation?

Check the attribution confidence on that row first. A pause driven by a channel with weak evidence is often a tracking problem rather than a performance one, and cutting spend will not fix it.

Related

• How MarginOS calculates channel LTGP:CAC — the headline ratio in this table.
• How MarginOS assigns each order to a channel — what every row here depends on.
• How MarginOS resolves blended CAC — the cost side.
• Read the Command Center — the business-level view this drills down from.