Stop Optimizing for ROAS. Optimize for CAC Payback.

ROAS is the most beloved and most misleading metric in paid media. Every platform reports it, every agency celebrates it, and most founders use it as the primary yardstick for ad spend. It's also deeply, structurally misleading for any business with repeat purchase, lifetime value, or product margin that varies by SKU — which is most businesses.

This piece makes the case for moving off ROAS as your primary lens and onto new-customer contribution margin and CAC payback windows. The change is not cosmetic. In our experience, clients who make this reporting switch typically reallocate budgets within 30 days and see bottom-line profitability improve within a quarter, sometimes without changing tactics at all.

What ROAS quietly hides

ROAS is revenue divided by ad spend. Simple, but every word hides a trap. 'Revenue' in ROAS calculations almost always means gross revenue, not net of returns, shipping, COGS, or discount codes. For a DTC brand running a 20% off code to acquire customers, a 3.5x ROAS is often a 1.2x gross-margin return — break-even at best. The founder reads 3.5x and thinks the campaign is printing money.

The second trap: ROAS treats first-time and repeat purchases the same. A $100 order from a new customer is worth about $40 in contribution margin after COGS, shipping, payment processing, and the ad cost that acquired them. A $100 repeat purchase from an existing customer is worth closer to $70 because the acquisition cost was amortized long ago. Meta's in-platform ROAS mashes these together, overstating performance by 20–40% depending on repeat rate.

The third trap: attribution windows. Meta's default 7-day-click, 1-day-view window credits the platform for conversions a lot of which would have happened anyway. Google's last-click model does the same. A 5x reported ROAS can be a 2.5x true incremental ROAS after adjusting for baseline organic demand. Optimizing hard for a number that overstates reality by 2x is how companies spend themselves into unprofitability while their dashboard still looks healthy.

The metric we actually watch — new-customer contribution margin

New-customer contribution margin is revenue from first-time buyers minus COGS, minus shipping and processing, minus ad cost. If it's positive on day zero, you have a profitable acquisition engine that doesn't require LTV assumptions to work. If it's negative, you're betting the company on future repeat purchases — which requires honest retention data you probably don't have yet.

The math is straightforward. For a DTC brand with 35% gross margin, a $50 first-order AOV, and a $20 blended CAC, new-customer contribution margin is $50 × 0.35 − $20 = −$2.50 per new customer. Break-even requires either a higher first-order AOV (upsells at checkout), lower CAC (better creative or channel mix), or real LTV from repeat purchases. The negative number doesn't mean the business is broken — it means you need to clearly see what has to be true for it to become profitable.

For SaaS, the equivalent is customer lifetime contribution margin minus CAC divided by CAC. Anything under 3x long-term is marginal. Anything under 1x is a company that won't survive.

Payback windows that hold up across cycles

The CAC payback window is the number of months it takes for a newly acquired customer's contribution margin to equal the cost of acquiring them. It's the single most important health metric for a paid acquisition engine because it determines how long your capital is tied up before you can redeploy it.

Reasonable benchmarks by category: under 3 months for DTC with strong recurring purchase (beauty, supplements, consumables). Under 6 months for DTC with moderate repeat (apparel, home goods). Under 12 months for SaaS with healthy expansion revenue. Under 18 months for complex B2B enterprise deals. Anything longer is a high-risk business model — you're betting the company on repeat behavior you can't validate fast enough, and capital markets won't forgive you if the model turns out to be wrong.

The reporting change that drives the behavior change

We stopped showing ROAS on monthly client decks about 18 months ago. What replaced it: new-customer payback curves by cohort, blended MER, contribution margin by channel, and an honest incrementality estimate once per quarter. The effect was immediate — clients started asking better questions. Instead of 'why is Meta ROAS down this month,' they'd ask 'why did cohort payback extend from 4.2 to 5.1 months, and which channels contributed to that.'

Better questions lead to better decisions. One DTC client cut Meta spend 30% and reinvested it in email and SMS based on the new reporting, which shortened payback by almost two months and lifted overall profitability 15% within a quarter. Same team, same product, same budget — just a better metric.

Key takeaways

• ROAS hides COGS, discounts, repeat-purchase blending, and attribution inflation. Reported ROAS typically overstates reality by 20–40%.
• Track new-customer contribution margin instead — revenue minus COGS minus ad cost. Positive = profitable engine.
• Target payback windows: DTC under 3–6 months, SaaS under 12, B2B under 18. Longer is a high-risk model.
• Changing the report changes the questions. Questions change decisions. Decisions change profitability.