Most DTC paid media metrics lie to you, and the dashboards that show them are designed to lie a specific direction. Platforms over-credit their own contribution by 15% to 40% through view-through attribution and double-counting between Meta and Google. A single metric, especially Platform ROAS, will quietly hide a structural unprofitability under a green number.
The fix is not better attribution software. It is reading the metrics as a three-layer stack, in order, with the right question at each layer.
The metric that matters depends on what you are scaling
DTC paid media metrics work in three layers. Platform (Hook Rate, CTR, CPM) tells you if the creative resonates. Conversion (CVR, Platform CPA, Platform ROAS) tells you if the funnel converts. Business (MER, Blended CAC, Contribution Margin, CAC Payback) tells you if the company is profitable. Diagnose top-down. Decide spend bottom-up. A 4x platform ROAS is meaningless if Blended CAC exceeds what your CM2 can repay.
Capital efficiency expectations have tightened severely. Mature DTC brands now target a Contribution Margin of 35% to 60% and a CAC Payback under 90 to 120 days, and the industry's blended CAC has climbed 40% to 60% since 2023. Operating on platform-level metrics in that environment is how brands scale into insolvency while celebrating ROAS.
The three-layer stack at a glance
Every metric on a paid social dashboard belongs in one of three layers. Each layer answers a different question and triggers a different decision.
| Layer | What it diagnoses | Lead metric | Decision it drives |
|---|---|---|---|
| Platform | Attention and auction dynamics | Hook Rate | Kill or keep the creative |
| Conversion | Friction and platform efficiency | CVR + Platform CPA | Fix the LP or change the offer |
| Business | Unit economics | MER + Contribution Margin + CAC Payback | Scale, hold, or cut spend |
The rule is simple and almost universally violated: diagnose from the top of the funnel down, but make scaling decisions from the unit economics up. A campaign that breaks at the platform layer cannot be fixed by tweaking budget. A campaign that breaks at the business layer cannot be saved by better creative.
Layer 1: platform metrics (attention and auction)
Layer one lives entirely inside the ad platform and exists to diagnose the creative. With Meta Advantage+ Shopping Campaigns now eating roughly 62% of ecom conversion spend, and the Andromeda retrieval engine reading the creative itself to predict who will buy, you no longer control targeting. You control what the algorithm sees. That makes performance creative the lever, and the platform layer the way you read whether the lever is moving.
Hook rate, the entry metric
Hook Rate is 3-second video plays divided by impressions on Meta, 2-second on TikTok. It is the single most important diagnostic on the platform layer.
Benchmark bands:
- Meta: 25% to 30% is solid, 40%+ is elite, anything under 15% to 20% is failing the auction.
- TikTok: 30% to 35% baseline, 40%+ top quartile. Native sound and faster scroll velocity push these higher than Meta for equivalent creative.
- UGC and direct-to-camera founder testimonials: routinely 60% to 70%, and pull CPAs 20% to 30% below polished studio cuts.
If Hook Rate is under 20%, every metric downstream is mathematically irrelevant. People are scrolling past your pitch before it starts. Nothing about your landing page or offer changes that.
Hold rate, CTR, CPM, CPC
The rest of the platform layer is a sequential diagnostic, not a target list.
| Metric | Meta band | TikTok band | What a miss means |
|---|---|---|---|
| Hold Rate | 40% to 50% of hooked viewers | similar | Body of the ad fails to deliver the hook's promise |
| CTR | 1.5% to 2.8% | 0.5% to 1.5% | Offer or CTA lacks urgency; on Meta, under 1.5% triggers CPM penalty |
| CPM | $18 to $45 | $4 to $13 | Account is fragmented, creative is stale, or hook is failing |
| CPC | $0.70 to $1.35 | $0.30 to $1.50 | Downstream effect of bad CPM or weak CTR |
The interplay matters more than the absolute numbers. Bad creative drives a low Hook Rate and low CTR, which the algorithm punishes with higher CPM, which inflates CPC, which inflates CPA. The cure is upstream. A fragmented account with bloated ad set counts and stale creative drifts to the top of the CPM band; a clean Advantage+ build with active creative rotation drifts to the bottom of the same band. The architectural fix lives in account structure.
How to read the platform layer
Read it sequentially, top of the funnel down.
- Hook fails: rewrite the first three seconds. Nothing else.
- Hook fine, Hold fails: the body of the ad does not deliver on the opening promise.
- Both fine, CTR fails: the offer or CTA is the weak link.
This is the ladder you climb before touching the landing page. Full benchmark tables across verticals live on paid social benchmarks.
Layer 2: conversion metrics (friction and platform efficiency)
Once an ad earns the click, the burden shifts to the site and to the platform's attribution math. Both can mislead you. CVR is the highest-leverage metric in the middle of the funnel: a lift from 2% to 3% raises ROAS by roughly 50% without spending another dollar. Platform CPA and Platform ROAS are useful for ranking ad sets against each other, and dangerous when used as the basis for scaling decisions.
CVR is contextual, not universal
A 3% conversion rate is not a benchmark. It is an average across wildly different products and devices, and an average is the worst possible target.
Tier CVR by AOV first.
| Segment | Median CVR | Top 25% |
|---|---|---|
| AOV under $30 | 4.2% | 7.1% |
| AOV $30 to $100 | 2.9% | 5.0% |
| AOV $100 to $300 | 2.1% | 3.8% |
| AOV $300+ | 1.4% | 2.6% |
| Mobile traffic | 1.8% to 2.87% | ~3% |
| Desktop traffic | 3.2% to 4.51% | 5.4%+ |
| Email (existing customers) | 4% to 8% | higher |
| Google Search (paid) | 2% to 5% | higher |
| Meta cold prospecting | 1.5% to 2.5% | retargeting 2% to 5% |
| TikTok paid social | 0.8% to 2.0% | higher |
The mobile gap deserves its own paragraph. Mobile drives roughly 84% of DTC traffic and converts at less than half the rate of desktop. That 42-point gap is the largest unfixed revenue leak in most stacks. Fashion is the worst offender: 78% mobile traffic, 1.2% mobile CVR against 1.9% desktop CVR.
A 1.2% CVR on a $250 AOV jewelry brand is strong. A 3% CVR on a $20 snack brand is weak. Stop comparing yourself to a global average.
Platform CPA and Platform ROAS, and why they lie
The numbers on the dashboard are generated by a system designed to claim maximum credit.
- Meta median CPA: roughly $38.17 across DTC, with vertical ranges from $30 (baby, lifestyle) to nearly $50 (electronics, travel).
- TikTok median CPA: roughly $32.74 for ecommerce.
- Platform ROAS, Meta: 1.86x to 2.19x blended.
- Platform ROAS, TikTok: 1.41x average, climbing to 2.25x with Value Optimization.
The structural problem: platforms over-credit by 15% to 40% through view-through attribution and double-counting. Run Meta and Google Performance Max simultaneously and both will routinely claim 100% of the same purchase. The sum of platform-reported revenue often exceeds actual Shopify revenue by 30% to 60%.
Use these for relative comparison between ad sets in the same account. Do not use them as the basis for scaling decisions. The attribution deep dive lives on marketing attribution, and the causal question (does this channel create demand or just capture it) belongs to incrementality testing.
Layer 3: business metrics (unit economics)
This is the only layer that touches the bank account, and the only layer that should drive a scaling decision. Capital efficiency expectations have tightened severely post-ZIRP. Top-line GMV growth on broken unit economics is now a liability, not a win, and the WACC for consumer retail has climbed to 7.5% to 11.0%.
MER, the macro truth metric
MER is Total Company Revenue divided by Total Marketing Spend, attribution-agnostic. It does not care which platform claims credit. It only cares whether the marketing machine, as a whole, is profitable.
- Healthy scaling: 3.0x to 5.0x.
- Median: about 4.0x.
- Danger zone: below 2.0x is structurally unprofitable for most margin profiles.
- Apparel: 2.1x to 3.4x.
- Supplements / wellness: 3.0x to 5.5x.
- Furniture / home goods: 1.5x to 2.5x, justified by AOV.
The scaling pattern that scares brands (and shouldn't): you increase Meta prospecting spend 30% to 50%, Meta-reported ROAS drops 10% to 20%, and MER stays flat or rises. That is not a problem. That is profitable scaling, because the new spend is generating demand that Google PMax and email harvest. Pausing Meta prospecting because its in-platform ROAS dropped is one of the most expensive mistakes in DTC. The mechanic and the diagnostic test for this pattern is the entire point of MER vs ROAS.
Blended CAC vs Paid CAC, the fully-loaded number
The CPA on your Meta dashboard is not your CAC. It is one input to it.
- Ecommerce blended CAC median: $68 to $87, up 40% to 60% since 2023.
- Paid CAC vs Blended CAC: Paid runs 2.4x to 3.1x higher than blended for most brands.
- The hidden cost stack: agency retainers, creative production, attribution software, and marketing salaries belong in the numerator. A $40 platform CPA routinely materializes as an $80 to $120 fully-loaded Blended CAC once those are loaded in.
- The returns adjustment: a $75 apparel headline CAC against a 26% return rate becomes a $101 True CAC. Swimwear and lingerie return rates run 30% to 50%.
If you are scaling on platform CPA without these loads, you are scaling into a deficit you cannot see on the dashboard. The full breakdown of the calculation and the payback formula is on customer acquisition cost.
Contribution margin and breakeven ROAS
Your CM2, the contribution margin available before ad spend, sets the absolute ceiling on what you can pay to acquire a customer. The math is unforgiving.
Breakeven ROAS = 1 รท CM2%
| CM2 | Breakeven ROAS | Implication |
|---|---|---|
| 60% (premium skincare, supplements) | 1.67x | A "3x target" leaves massive profitable scale on the table |
| 40% (standard apparel) | 2.50x | Universal 3x target works |
| 20% (thin-margin / dropship) | 5.00x | A 3x target is bleeding cash on every order |
A single ROAS target applied across every SKU simultaneously starves the high-margin product (artificially capping its scale) and drains cash on the low-margin product (scaling unprofitable volume). The full CM1 / CM2 / CM3 ladder, the POAS playbook, and how to feed margin into Meta and Google smart bidding via custom conversion values lives on contribution margin.
CAC payback period, the cash-flow metric
CAC Payback is Blended CAC divided by Monthly Contribution Margin per Customer. It tells you how fast acquired cash comes back, which determines how fast you can reinvest.
- Ideal: under 90 days.
- Healthy: 90 to 120 days.
- Dangerous: beyond 120 days without bridge financing.
The vertical nuance matters more than the headline number. A DTC food and beverage brand with a $45 to $53 Blended CAC may be negative on the first order, with payback stretching to four or five months, and only viable through a high-retention subscription model. A furniture brand with a $120 Blended CAC may pay back on Day 0 because of an $800 AOV, even if the customer never buys again. Pair this metric with the LTV:CAC ratio (3:1 minimum floor) and customer lifetime value before you decide whether your payback is acceptable.
The order to optimize in
The point of three layers is not three to-do lists. It is one decision flow.
Read top-down to diagnose:
- Weak Hook Rate? The creative is broken. Stop optimizing anything else.
- Hook fine, weak CVR? The landing page or offer is broken.
- CVR fine, weak Platform ROAS? Bidding, audience, or attribution.
Read bottom-up to decide spend:
- What is MER doing across the whole business?
- What does CAC Payback say about reinvestment velocity?
- What does Contribution Margin say about the maximum CAC we can absorb?
Those three answers set the spend ceiling. Everything above is the explanation for why you are at that ceiling, not permission to push past it.
What to measure with (the tracking stack)
Platform dashboards alone will not get you there. The 2026 default is triangulation across three independent methods.
- Multi-touch attribution software for daily decisions. Triple Whale for Shopify-native brands running concentrated Meta and TikTok mixes (pricing scales with GMV, starting around $1,490/year). Northbeam for heavier multi-channel operations using ML and first-party data ingestion (starting around $1,000/month). ProfitMetrics for brands focused on injecting CM2 directly into platform bidding (POAS). Cometly for brands syncing into Snowflake or BigQuery (Core plan $750/month).
- Marketing Mix Modeling for macro budget. MMM uses aggregate regression on spend, seasonality, and promotions to isolate incremental contribution per channel without relying on user-level cookies. The 2026 versions are AI-driven and updated in near-real-time.
- Incrementality and post-purchase surveys for ground truth. Geo-lift tests separate correlation from causation. "How did you hear about us?" survey responses routinely correct the 30% to 50% over-crediting that bottom-funnel channels (branded search, retargeting) inherit from last-click models.
The prerequisite that makes any of this accurate is server-side tracking through Conversions API, which bypasses browser-level signal loss and lets you push profit data, not just revenue, back to the algorithm. Full attribution architecture sits on attribution, and the causal validation step on incrementality testing.
Common ways brands read this stack wrong
A short, opinionated list. Every one of these is a real pattern we have seen kill profitable scale.
- Scaling on Platform ROAS while MER stays flat. You are paying the platform for credit, not the business for customers.
- One ROAS target across every SKU. Different CM2 profiles require different targets. A universal 3x simultaneously suffocates high-margin scale and bleeds low-margin spend.
- Optimizing CPA without loading agency, software, salaries, and creative. Paid CAC understates Blended CAC by 2x to 3x.
- Ignoring the mobile CVR gap. 84% of traffic, half the conversion rate. This is the largest unfixed leak in most DTC stacks, and almost no team is auditing for it.
- Treating Hook Rate as a vanity creative metric. It is the input to CPM, which is the input to CPC, which is the input to CPA. The auction reads it before you do.
- Killing prospecting because in-platform ROAS dropped while blended MER rose. You just paused the engine that was generating the demand the other channels were harvesting.
Where to go from here
Pick the layer that the diagnosis points at, then read the spoke.
- If the bottleneck is creative, route to performance creative and the testing framework.
- If it is structure, route to account structure and scaling paid social.
- If it is measurement, route to attribution and server-side tracking.
- If it is unit economics, route to contribution margin and LTV:CAC.
Get the diagnosis, not another dashboard
Most DTC brands do not need a fourth attribution tool. They need an outside read of the full stack against the actual P&L, so the next dollar of scale lands on the layer that is actually broken.
That is the entire point of a paid media audit: walk the three layers, anchor them to your contribution margin and payback, and tell you which layer to fix first. For brands whose audit pinpoints creative as the bottleneck, the next step is a performance creative agency engagement built to ship the volume of testing the Andromeda era requires.
Stop scaling on metrics that don't pay the bills.