Reading your Meta funnel: how to find the exact step that is leaking

Stop reading ROAS as one number. Read your Meta funnel as a chain of ratios, walk it top to bottom, and find the first step where the number falls off a cliff relative to the steps around it. That first big drop is where the money leaks, and it is the only step you should fix right now.

Most buyers do the opposite. ROAS dips, so they pause the ad, blame the creative, and rebuild. Sometimes the creative is fine and the landing page is slow. Sometimes the hook never landed and everything downstream was doomed. ROAS cannot tell you which, because ROAS is the whole chain multiplied together. One weak link drags the final number down, and you cannot see the link from the product.

The fix is to stop looking at the outcome and start looking at the steps.

The funnel is a chain of ratios

Every purchase passes through the same gates. Impression, then a scroll stop, then held attention, then a click, then a page load, then an add to cart, then a started checkout, then a purchase. Each gate has a ratio: how many people made it through divided by how many arrived. Read them in order.

The trick is not to judge any ratio against a benchmark you read somewhere. Judge it against its own neighbours. A step is leaking when its pass-through drops hard while the steps on either side look normal. That relative cliff is the signal. A number that is a little soft everywhere is a different problem than a number that is healthy, healthy, then collapses at one gate.

Walk down until you hit the first cliff. Stop there. Do not keep reading, do not fix three things at once. The first leak distorts everything below it, so you cannot trust the lower ratios until the top one is sealed. Fix the first cliff, let it run, then read the chain again.

Read the steps in order

Here is the chain, the ratio at each step, and what a weak reading usually means. The numbers are yours to pull day-wise. The diagnosis is what the shape tells you.

| Step | Ratio | A weak reading here usually points to |

|------|-------|----------------------------------------|

| Scroll stop | 3-second video views / impressions (thumb-stop rate) | The hook. The first frame, the opening line, the visual pattern. People are not stopping. |

| Held attention | thruplays / 3-second views (hold rate) | The creative body. The hook worked but the middle sags. Pacing, story, or proof is losing them. |

| The click | outbound CTR (clicks / impressions) | The ask. They watched but did not act. Weak call to action, no reason to click, offer not clear in the creative. |

| Page arrival | landing page view rate (landing page views / clicks) | Speed or relevance. The page is slow, or the click promised one thing and the page shows another. |

| Add to cart | add-to-cart rate (carts / landing page views) | The offer or the page. Price, product clarity, trust, or a page that does not sell once they arrive. |

| Checkout start | checkout rate (checkouts / carts) | Friction. Shipping shock, forced account creation, a slow or confusing cart. |

| Purchase | purchase rate (purchases / checkouts) | Trust and friction at the last step. Payment options, unexpected costs, a checkout that breaks. |

Read it as a story. A strong thumb-stop and a weak hold means the creative earned the stop and wasted it, so you cut the body, not the hook. A strong click and a weak landing page view rate means people wanted in and the page failed to load or matched nothing they were promised, so you fix the page, not the ad. A healthy add to cart and a weak checkout means the product sold and the checkout scared them off, so you look at shipping and forms, not the creative at all.

Fix the step, not the whole ad

This is the part that saves money. When you know the leak is at one gate, you change one thing. A hook problem gets new opening frames, not a new landing page. A body problem gets a re-edit of the middle, not a fresh offer. A checkout problem gets a friction pass, and the creative you were about to kill keeps running because it was never the issue.

Killing a whole ad over a checkout leak throws away a working hook and a working page. Rebuilding a landing page over a hook leak wastes a week and fixes nothing. The chain tells you where to aim so you spend effort on the one step that is costing you.

Two checks keep you honest. Before you act, confirm the top-line number you are reacting to is the right one to trust, because a falling ROAS and a falling contribution margin send you to different places. Start with which top-line metric to trust first. And when a step that used to be strong slowly weakens over days, ask whether the creative is tired or the leak is structural, because those need opposite fixes. Read is it fatigue or a real leak before you rebuild.

Make it a daily habit

The chain only works if you read it often and read it day-wise. Leaks move. A hook that stopped the scroll last week goes stale, a checkout that was clean breaks after a theme update, a page that was fast slows down when traffic climbs. If you only look at ROAS, you see the damage a week after it started and you cannot tell which gate moved.

This is the reading AdBrain does for you every day. It walks the chain for each ad, ratio by ratio, points to the first step that fell off a cliff, and tells you what that step usually means, so you open the account already knowing where to look. It does not touch your ads. It hands you the diagnosis and the reason, and you make the call.

Start tomorrow. Pull the ratios in order, find the first cliff, fix that one step, and let the rest of the chain tell you the truth once the top is sealed.