ROAS Thresholds by Stage: Kill, Optimize, or Scale
I get this question every week from founders spending anywhere from $5K to $500K a month: "Is a 2.1 ROAS good?" Wrong question. The right one is how to know if ROAS is good for my brand at this specific stage. A 2.1 ROAS can be a disaster or a home run depending on your margins, your CAC payback, and whether you're launching a brand or scaling one. I've run this call on a $2.2B infrastructure project's demand gen budget and on a DTC brand I took from $100K to $3M. The math is different every time. The framework isn't.
What the problem actually looks like
Founders come to me with a dashboard number and no context. "Meta says 1.8 ROAS." They don't know their contribution margin. They don't know if that 1.8 includes new customer acquisition only or blended with retargeting. They panic at a 2.0 that's actually profitable, or they keep scaling a 4.0 that's quietly cannibalizing organic and email revenue.
The tell is always the same: someone asks "is this good" without being able to answer "good for what decision." Kill it? Keep spending the same? Pour more in? Those are three different thresholds, and most people are using one number for all three.
Why it happens
ROAS is a platform-reported metric built to keep you spending on that platform. It's not a profitability metric. It ignores your cost of goods, your shipping, your returns rate, your fixed costs, and your payback period. A 3.0 ROAS on a product with 20% margin loses money. A 1.5 ROAS on a product with 70% margin and high repeat rate can be excellent.
The second reason is stage blindness. Early stage, you're paying a premium to learn who your customer is and build signal in the algorithm. Later stage, you have retargeting pools, lookalikes, and brand search lift doing work for free. Judging a launch campaign against a mature campaign's ROAS is comparing a rookie to a ten-year vet.
Third, people track ROAS on a 1-day or 7-day click window and treat it as gospel, when the real payback often shows up 30-60 days out through repeat purchase. Impatience kills good campaigns constantly.
What we do about it
Here's the actual threshold system I use, broken by stage. These aren't universal laws, they're operator heuristics that have held up across a lot of accounts.
Stage 1: Launch / Testing (0-90 days, under $20K/month spend)
- Kill: Below 1.0 ROAS after $1,500-2,000 spend per ad set with no upward trend over 10 days. You've bought enough data to know.
- Optimize: 1.0 to 1.8 ROAS. This is normal at launch. Don't scale, don't kill. Fix creative, tighten audience, test offers.
- Scale: Above 1.8 ROAS with at least 3 consecutive days of stability. Start increasing budget 20% every 2-3 days, not overnight.
Stage 2: Growth ($20K-$150K/month spend)
- Kill: Below your breakeven ROAS. Calculate it: 1 / gross margin. If margin is 55%, breakeven is 1.82. Anything consistently under that for 2+ weeks, kill or fully rebuild the campaign.
- Optimize: Breakeven up to breakeven + 0.5. At 55% margin that's 1.82 to 2.3. You're profitable but not efficient. Push creative refresh weekly, test new angles, watch frequency.
- Scale: Above breakeven + 0.5, sustained for 7+ days. This is where I'd move budget up 15-25% weekly and start layering in new channels.
Stage 3: Mature / Scaled (over $150K/month)
- Kill: Below breakeven ROAS for 10+ days despite creative rotation. At this size, a stalled campaign is bleeding real cash fast, not just wasting test budget.
- Optimize: Breakeven to breakeven + 0.3. At scale, small ROAS bands represent huge dollar swings, so this range gets tight scrutiny weekly, not monthly.
- Scale: Above breakeven + 0.3 with stable CAC and no drop in blended MER (marketing efficiency ratio, total revenue divided by total spend). Never scale on platform ROAS alone at this stage. Check blended numbers or you'll scale yourself into a cash crunch.
On the DTC brand I scaled from $100K to $3M, our breakeven ROAS was 1.65 (61% gross margin). We killed anything under 1.4 within two weeks. We scaled hard anything over 2.2. Everything in between got creative testing, never budget increases. That discipline is what let us compound instead of chasing ghosts.
What it costs to ignore
I've seen founders scale a 3.5 ROAS campaign on a 25% margin product and lose money every month while celebrating "record ROAS." I've seen others kill a 1.6 ROAS campaign that would have paid back at 3.2 lifetime value over 90 days, because they only looked at day-7 numbers.
On the infrastructure project, we ran a lead gen campaign with a "ROAS" (really cost per qualified lead translated to pipeline value) that looked terrible at day 30 and excellent at day 120, because the sales cycle was long. If we'd killed it at day 30 using a DTC mindset, we'd have thrown away $400K in pipeline.
The cost of ignoring stage-specific thresholds isn't abstract. It's wasted spend on the kill side and missed compounding growth on the scale side. Both are expensive. Most brands lose more from premature killing than from slow bleeding, because they never let campaigns reach statistical or temporal maturity.
Most common mistake
The single biggest mistake: using one flat ROAS target across every stage and every campaign type. A 2.0 target sounds simple. It's also meaningless without margin, payback window, and stage attached to it. If your rule doesn't reference your gross margin, it isn't a rule, it's a guess.
FAQ
What ROAS is considered good for e-commerce?
There's no universal number. Calculate your breakeven ROAS as 1 divided by your gross margin percentage. A 50% margin brand breaks even at 2.0. Good means sustainably above that, not any specific figure you saw on a benchmark blog.
How long should I wait before killing a low ROAS campaign?
At launch stage, give it $1,500-2,000 in spend or 10 days, whichever comes first. At scale, 10-14 days is enough because sample size is already large. Longer sales cycles (B2B, high-ticket) need 60-90 day windows tied to actual close data, not platform attribution.
Practical takeaway
Write your breakeven ROAS on a sticky note before you open your ads dashboard again. Everything else, kill, optimize, scale, gets measured against that one number, adjusted for your current stage. Stop asking if a ROAS is good in general. Ask if it's good for the math of your business, this month, at this spend level.