The Ops Loop

Find the leak in your funnel in 90 seconds.

Four queries. Four benchmarks. They tell you whether your problem is traffic or the page — before you raise your ad budget.

No signup. No email. Copy them and run them.

Query 1

Where is the traffic coming from?

FROM sessions SHOW sessions GROUP BY referrer_source ORDER BY sessions DESC SINCE -30d UNTIL today

What it answers: whether your ads are delivering at all.

If most sessions come from paid social or paid search, the top of the funnel works. Ruling traffic out is the first step — and the one almost nobody takes before raising the budget.

Query 2

Where exactly do people drop?

FROM sessions SHOW sessions, sessions_with_cart_additions, sessions_that_reached_checkout, sessions_that_completed_checkout, conversion_rate SINCE -30d UNTIL today

The number that matters: cart additions ÷ sessions.

If it sits far below the normal range, your problem is the page, not the advertising. No amount of extra traffic fixes a page that does not convince.

Query 3

What happens after the cart?

Same data as Query 2, read separately — because the causes are different:

checkout ÷ cart drops → friction in the cart itself.
paid ÷ checkout drops → usually shipping cost revealed late, missing payment methods, or forcing account creation.

Query 4

Is it worth fixing first?

FROM sales SHOW orders, gross_sales, net_sales, total_sales, average_order_value SINCE -30d UNTIL today

Then run the arithmetic that decides where your next hour goes:

projected orders = sessions
  × target add-to-cart rate
  × (checkout ÷ cart)      ← leave as-is
  × (paid ÷ checkout)      ← leave as-is

Change only the add-to-cart rate. Leave every downstream rate exactly as broken as it is today. What comes out is how many orders you gain without spending one extra dollar on ads.

The benchmarks

Metric	Normal	Alarm
Add to cart	5–8%	under 3%
Cart → checkout	60–80%	under 50%
Checkout → paid	45–65%	under 35%
Overall conversion	1.5–2.5%	under 0.8%

Do not chase perfect numbers. Find which of the four sits furthest from its range — that is the one you work on first. That is the whole prioritization.

One honest caveat. Under roughly 30 orders in the window, these numbers give you direction, not proof. Use them to decide where to work. Do not use them to declare that something worked.

Not on Shopify?

The queries are ShopifyQL, but the same four numbers exist everywhere:

WooCommerce	Analytics → Conversions
BigCommerce	Analytics → Merchandising
Google Analytics 4	Monetization → Purchase journey
Anything else	GA4 funnel exploration with the four events

Want the number found for you?

If your funnel comes back ugly and you would rather have it fixed than explained, I run a paid teardown: the four numbers pulled from your store, ranked by how much money each leak is costing you, with the specific fix for the worst one.

Book a funnel teardown