Your Best Hook by Hook-Rate Is Your Worst by Margin

Your top creative by hook-rate is doing exactly what you rewarded it for: stopping thumbs. But stopping thumbs and earning margin are two different jobs, and the ad that’s best at the first is frequently quietly the worst at the second. Rank your creative tests by upper-funnel attention and you will systematically promote your least profitable ads — then wonder why a “winning” account keeps shrinking your contribution line.

For the surrounding account decisions, compare Hook Rate vs Hold Rate: What They Mean and the Traps and use Financial Services Meta Ads: Policy and Approval Guide as the next diagnostic.

The metric you reward is the audience you get

A creative is not a passive asset that sits inside your targeting. It is targeting. Drop two ads into the same ad set with identical audience settings and the delivery system will still serve them to different sub-populations, because it hunts for the least expensive people likely to take the action you optimized for. The hook does a lot of that sorting before anyone clicks.

A loud pattern-interrupt — fast cuts, a shock open, a curiosity gap — lowers your CPM and lifts your hook-rate because it appeals broadly. Broad appeal is the problem. It pulls in scrollers, browsers, and bargain-hunters alongside genuine buyers. The creative that earns a 38% thumbstop is frequently the one casting the widest, least-qualified net.

This is Goodhart’s law in an ad account: the moment hook-rate becomes the target you select on, it stops measuring creative quality and starts measuring how well a thumbnail interrupts. Those are not the same thing, and the gap between them is paid for out of your margin.

Why the hook-rate winner bleeds margin

Walk the chain a high-hook creative actually produces:

• Cheaper, shallower attention. High hook-rate, lower hold rate. People stop but don’t absorb the offer, so click intent is thin.
• Lower commercial intent. Click-through can look fine while add-to-cart rate sags. The cohort was curious, not in-market.
• Worse basket economics. When the curious do buy, they skew toward the entry SKU, lean harder on discount codes, and convert at a lower average order value.
• Weaker durability. Impulse-led, discount-led purchases return and refund at higher rates and retain worse — so even the revenue you booked overstates what you keep.

Every one of those steps is a leak below the platform metrics many teams grade on. Hook-rate, thru-play, and even outbound CTR all live above the point where unit economics are decided. Contribution margin — revenue minus COGS, fulfillment, payment processing, and a returns reserve — is decided at the bottom. A creative can win every metric above the fold and still generate less margin per impression than the “boring” ad it beat.

Rank by contribution margin per impression

The honest unit for ranking creative is contribution margin produced per unit of delivery — call it CM per thousand impressions, or CM per unit of spend. It captures the whole chain a creative drags behind it: the audience it selects, the basket it sells, and the returns it leaves you holding.

Consider two ads in the same test. The numbers below are illustrative, not benchmarks — the point is the shape, not the values:

Signal	Creative A (hook winner)	Creative B (margin winner)
Hook rate (3s)	38%	22%
Outbound CTR	2.1%	1.4%
Add-to-cart rate	4%	7%
Purchase rate	1.1%	1.9%
AOV index	0.85	1.15
Return rate	elevated	low
CM per 1,000 impressions	1.0x (baseline)	~1.6x

Creative A wins the metrics a creative review in many cases celebrates. Creative B wins the only one that pays salaries. If you’d killed B on hook-rate in week one — which is exactly what a hook-rate leaderboard tells you to do — you’d have scaled the ad that erodes your contribution line and buried the one that funds it.

Build a contribution-margin layer into creative testing

The fix is structural: creative testing needs a contribution-margin layer underneath the platform dashboard, not instead of it. Practically:

1. Define the margin stack once, explicitly. Revenue − COGS − fulfillment − processing − returns reserve = CM2. Subtract ad spend for contribution after marketing. Agree on this with finance so “profitable” means one thing.
2. Carry creative identity downstream. The hard part is joining post-purchase reality — AOV, refund rate, discount usage, new-vs-returning, the SKU actually sold — back to the creative that drove the order. Most stacks drop creative ID at checkout; instrument so it survives.
3. Grade on the bottom of the ladder, not the top. Read the funnel as a ladder of economic quality: hook (attention) → hold (interest) → CTR (intent) → landing-page view (qualified intent) → add-to-cart (commercial intent) → purchase → AOV and margin mix → return rate (durability). Rank creative on the last three rungs. Use the top rungs as diagnostics, not verdicts.
4. Watch the rank gap as an early signal. The most useful number you’re not tracking is the delta between a creative’s hook-rate rank and its add-to-cart rank. A creative that’s top-three on hook and bottom-half on add-to-cart is the classic cheap-attention trap — flag it before it scales, not after.

This is the kind of read-only diagnostic Bach AI is built to surface: it reads the account end-to-end, ties downstream margin back to each creative, and proposes the reallocation — but it stays read-only until you approve the move.

Don’t kill creative on the wrong signal — or too early

Two failure modes sit on either side of this.

The first is judging too fast. The delivery system needs enough recent optimization-event signal before its read on a creative stabilizes; as a rough planning range, teams frequently wait until a creative has gathered on the order of tens of purchases before trusting its numbers. Below that, you’re ranking noise — and noise favors the loud hook, because attention metrics fill up first.

The second is judging on the wrong layer. Hook-rate is a fine diagnostic: a low one in many cases means the creative isn’t being seen, which is worth fixing. It is a terrible verdict, because being seen and being bought are different outcomes with different economics. Let hook-rate tell you whether a creative got a fair hearing. Let contribution margin tell you whether it earned its place.

The takeaway

Attention is cheap, and the auction will happily sell you more of it. Margin is scarce, and it’s decided below every metric on the default dashboard. Stop ranking creative by who stops the most thumbs and start ranking by contribution margin per impression — the ad that selects the right audience, sells the right basket, and keeps the sale. Your best hook is frequently your worst buy. Grade the buy.