ATTN.
← Back to Blog

2026-03-02

Blended ROAS vs. New Customer ROAS: Why the Distinction Matters

Blended ROAS vs. New Customer ROAS: Why the Distinction Matters

Blended ROAS vs. New Customer ROAS: Why the Distinction Matters for DTC Brands

Every DTC brand tracks ROAS. It's the first number you check in the morning and the last one you argue about before bed. But here's the problem: if you're only looking at blended ROAS, you're flying blind.

Blended ROAS combines all revenue — new customers and returning customers — into one number. It's simple. It feels good. And it lies to you constantly.

The Math That Misleads

Say you spent $50K on Meta last month and generated $200K in revenue. That's a 4x blended ROAS. Looks great, right?

But dig deeper:

  • $120K came from returning customers who would have purchased anyway (email, organic, direct)
  • $80K came from actual new customer acquisition
  • Your real new customer ROAS is 1.6x, not 4x

That 4x number made you feel like a genius. The 1.6x number tells you the truth: your acquisition efficiency is barely breaking even, and you're subsidizing it with retention revenue you didn't earn through paid media.

Why This Distinction Matters

Blended ROAS inflates performance because it takes credit for revenue that wasn't driven by the ad spend. Returning customers have higher conversion rates, higher AOV, and lower cost to convert — they already know and trust your brand.

When you blend them in:

  • Your ROAS looks artificially high
  • You think you can scale spend without consequences
  • You increase budgets based on false signal
  • Margins deteriorate as you scale, and you can't figure out why

New customer ROAS isolates the actual cost of growth. It tells you what you're paying to bring in someone who has never purchased before — the true engine of business expansion.

How to Calculate Each

Blended ROAS = Total Revenue / Total Ad Spend

New Customer ROAS = New Customer Revenue / Total Ad Spend (or acquisition-specific spend if you can isolate it)

The challenge is attribution. Platforms like Meta will claim credit for returning customers who saw an ad before purchasing. You need either:

  • Platform-level new customer exclusions (Meta's existing customer lists)
  • A third-party attribution tool (Triple Whale, Northbeam, Rockerbox)
  • Shopify's built-in new vs. returning customer data cross-referenced with ad spend

The Benchmark Gap

In our experience managing $500K+/month across DTC brands:

  • Average blended ROAS: 3-5x
  • Average new customer ROAS: 1.2-2.5x
  • The gap between the two: typically 40-60%

If your blended ROAS is 4x but your new customer ROAS is below 1.5x, you're not scaling — you're just retargeting your existing customer base more aggressively.

What Good Looks Like

For a healthy DTC brand:

  • New customer ROAS of 2x+ means your acquisition is profitable after COGS
  • New customer ROAS of 1.5-2x means you're breaking even on first purchase (acceptable if LTV is strong)
  • New customer ROAS below 1.5x means you're buying customers at a loss — only sustainable if your repeat purchase rate and LTV justify the upfront investment

The key is knowing your contribution margin. If your product has 70% margins, a 1.5x new customer ROAS still makes money. If you're at 40% margins, you need 2.5x+ just to break even.

How to Use Both Numbers

Don't abandon blended ROAS — it's still useful as a health check on overall marketing efficiency. But use new customer ROAS as your primary scaling signal:

  1. Set budgets based on new customer ROAS targets, not blended
  2. Build separate campaigns for prospecting (new) and retargeting (existing)
  3. Measure each independently — don't let retargeting revenue inflate your prospecting performance
  4. Use new customer ROAS to decide when to scale — if it holds above your target as you increase spend, you've found real signal

The Bottom Line

Blended ROAS tells you how your marketing portfolio performed. New customer ROAS tells you whether your business is actually growing. One measures the past. The other predicts the future.

If you can't tell the difference between the two in your reporting, you don't have a measurement problem — you have a strategy problem.