What Is a Good ROAS for Meta Ads? CPA and ROAS Percentiles from 82 DTC Brands (2026)
A good ROAS on Meta is one above your break-even; these percentiles show whether yours is typical. Across 79 DTC brands in the 90 days to 28 October 2026, the median Meta-attributed first-purchase ROAS was 1.53, the top quarter cleared 1.89 and 7.6% of brands reached 3.0. The median CPA was $63.80; the cheapest tenth paid under $29.33 a purchase, the dearest tenth over $160.41. Source: MHI Meta Benchmarks, DTC, October 2026 edition, read from MHI Media's warehouse of client Meta ad accounts on 29 October 2026, data settled to 28 October 2026.
These percentiles come from the Meta ad accounts of 82 DTC brands, read from MHI Media's warehouse of client ad accounts on 29 October 2026, data settled to 28 October 2026. Facebook and Instagram only. A brand is one client with all its Meta ad accounts added together, counted once in a window if it spent at least £3,000 there with at least 30 purchases: 79 brands over the 90 days from 31 July to 28 October 2026, 82 over the 12 months from 29 October 2025. The 10th, 25th, 50th, 75th and 90th percentiles are read across brands with linear interpolation; the median and p25 to p75 interquartile range follow the methodology on the benchmarks hub. ROAS is Meta-attributed first-purchase revenue divided by spend. USD at £1 = $1.3289; multiply by 0.7525 for GBP.
Metrics cover different dates. CPA, ROAS, spend, purchases and AOV run to 28 October 2026 in both windows. CTR, CPM, CPC and conversion rate exist for 8 January to 7 September 2026 only: 31 July to 7 September 2026 in the 90-day window (57 brands), 8 January to 7 September 2026 in the 12-month window (60 brands). Frequency (daily, campaign level) runs 8 January to 3 August 2026 (51 brands). Placement: 8 June to 13 July 2026 only, not used here. The dataset does not measure blended or lifetime revenue, repeat orders, profit or other channels; every ROAS below is first order.
What is a good ROAS for Meta ads?
A good ROAS for Meta ads is one above your break-even ROAS (1 divided by your contribution margin); the percentiles here tell you only whether your figure is typical. Across 79 DTC brands in the 90 days to 28 October 2026, the median first-purchase ROAS was 1.53. The break-even ROAS post walks through the calculation and the ROAS guide covers the definitions.
Which way is good:
- CPA, CPM and CPC: lower is better; p10 is the cheapest tenth of brands, p90 the dearest.
- ROAS, link CTR and conversion rate: higher is better; p90 is the best tenth, p10 the worst.
- AOV: a property of the product; higher leaves more room on CPA.
- Frequency: context only.
Every ROAS here is first order and Meta-attributed, so a 1.53 is a profit on a product people reorder and a loss on a thin-margin one-off purchase. Work out your break-even first, then use the tables to see how far from typical you are.
ROAS and CPA percentiles, 90 days to 28 October 2026
Over the 90 days to 28 October 2026, the middle half of 79 DTC brands sat between a 0.90 and 1.89 first-purchase ROAS and paid $41.76 to $98.54 per purchase; the medians were 1.53 and $63.80 (£48.01).
| Metric | n | p10 | p25 | Median | p75 | p90 |
|---|---|---|---|---|---|---|
| CPA | 79 | $29.33 | $41.76 | $63.80 (£48.01) | $98.54 | $160.41 |
| ROAS (first-purchase, Meta-attributed) | 79 | 0.66 | 0.90 | 1.53 | 1.89 | 2.38 |
| CTR (link) | 57 | 1.03% | 1.21% | 1.70% | 2.42% | 3.17% |
| CPM | 57 | $10.86 | $16.55 | $27.50 | $45.90 | $57.99 |
| CPC (link) | 57 | $0.81 | $1.10 | $1.63 | $2.17 | $2.58 |
| Conversion rate | 57 | 0.83% | 1.54% | 2.68% | 3.72% | 4.27% |
| AOV | 79 | $46.94 | $60.94 | $73.12 | $118.90 | $216.99 |
Source: MHI Meta Benchmarks, DTC, October 2026 edition (benchmark cuts, 29 October 2026).
The cheapest tenth of brands paid under $29.33 a purchase and the dearest tenth over $160.41. All three inputs spread between p10 and p90: CPM $10.86 to $57.99, CTR 1.03% to 3.17%, conversion rate 0.83% to 4.27%. I would check conversion rate first, because its fix sits on your site rather than in the auction.
ROAS and CPA percentiles, 12 months to 28 October 2026
Over the 12 months to 28 October 2026, the median first-purchase ROAS across 82 DTC brands was 1.49 and the median CPA $59.52 (£44.79), with the middle half between 0.94 and 2.02 on ROAS and $41.52 to $99.02 on CPA.
| Metric | n | p10 | p25 | Median | p75 | p90 |
|---|---|---|---|---|---|---|
| CPA | 82 | $27.56 | $41.52 | $59.52 (£44.79) | $99.02 | $163.26 |
| ROAS (first-purchase, Meta-attributed) | 82 | 0.66 | 0.94 | 1.49 | 2.02 | 2.47 |
| CTR (link) | 60 | 1.03% | 1.21% | 1.79% | 2.39% | 2.98% |
| CPM | 60 | $10.91 | $14.78 | $23.74 | $39.77 | $54.19 |
| CPC (link) | 60 | $0.70 | $1.03 | $1.40 | $1.96 | $2.78 |
| Conversion rate | 60 | 0.77% | 1.42% | 2.44% | 3.15% | 4.19% |
| Frequency | 51 | 1.30 | 1.38 | 1.50 | 1.68 | 1.92 |
| AOV | 82 | $43.29 | $59.04 | $74.38 | $116.62 | $237.40 |
Source: MHI Meta Benchmarks, DTC, October 2026 edition (benchmark cuts, 29 October 2026).
Use this table when your account is seasonal. The p75 ROAS is 2.02 here against 1.89 over 90 days and the median CPM $23.74 against $27.50, so the recent window ran dearer per thousand. The AOV benchmarks page splits CPA and ROAS by order value; AOV runs from $43.29 to $237.40 between p10 and p90.
How many brands clear a 1x, 2x or 3x ROAS?
Over the 90 days to 28 October 2026, 69.6% of 79 DTC brands cleared a 1.0 first-purchase ROAS, 50.6% cleared 1.5, 20.3% cleared 2.0 and 7.6% cleared 3.0.
| Threshold | 90 days (79 brands) | 12 months (82 brands) |
|---|---|---|
| ROAS at or above 1.0 | 69.6% | 70.7% |
| ROAS at or above 1.5 | 50.6% | 48.8% |
| ROAS at or above 2.0 | 20.3% | 25.6% |
| ROAS at or above 3.0 | 7.6% | 7.3% |
| CPA at or below $30 | 11.4% | 12.2% |
| CPA at or below $50 | 35.4% | 37.8% |
| CPA at or below $75 | 60.8% | 62.2% |
| CPA at or below $100 | 74.7% | 75.6% |
Source: MHI Meta Benchmarks, DTC, October 2026 edition (benchmark cuts, 29 October 2026).
This is the table I would send to anyone handed a 3x target. About one brand in thirteen reached it over 90 days, on Meta's own attribution. One in five cleared 2.0. About three in ten ran under 1.0, so they got back less first-order revenue than they spent; some are sound repeat-purchase businesses Meta cannot see. Set the target from your margin, then check here how rare it is.
Find your number
- Pick the window: 90 days for a current read, 12 months if seasonal.
- Use the account-level figure on Meta's attribution, first-order revenue for ROAS; blended ROAS and MER do not belong here.
- Find the two columns it sits between in Table C1 or C2, then the nearest threshold in Table C3.
| Your 90-day figure | Sits between | Reading |
|---|---|---|
| CPA $45 | p25 $41.76 and the median $63.80; 35.4% of brands at or under $50 | 25th to 35th percentile, roughly the best third |
| CPA $100 | Above p75 $98.54; 74.7% at or under $100 | The dearest quarter |
| ROAS 1.2 | p25 0.90 and the median 1.53; 50.6% clear 1.5 | Lower half, above the bottom quarter |
| ROAS 2.0 | Above p75 1.89; 20.3% at or above 2.0 | Top fifth |
| ROAS 3.0 | Above p90 2.38; 7.6% at or above 3.0 | Top decile |
Source: MHI Meta Benchmarks, DTC, October 2026 edition (benchmark cuts, 29 October 2026).
Your vertical moves the bar: the hub's vertical tables and the CPA by DTC vertical post put categories at different points on this scale.
What this means for your account
Your percentile band on CPM, CTR and conversion rate tells you where the purchase is being lost, and that decides what to fix first. Thresholds are from Table C1, 90 days to 28 October 2026.
- CPM above p75 ($45.90): narrow audiences, restricted placements or weak creative. Open targeting and placements, then test new concepts.
- Link CTR below p25 (1.21%): the ad is not earning the click. New hooks and first frames, more concepts a week.
- Conversion rate below p25 (1.54%): the problem is after the click: offer, price, landing page, ad-to-page match. Fix those before buying more traffic.
- All three near the median ($27.50, 1.70%, 2.68%) with ROAS under 1.0: order value. A $63.80 CPA on an AOV under the p25 of $60.94 is a first-order ROAS below 1. Bundles, AOV, or a measurable repeat model.
- CPA below p25 ($41.76) and ROAS above p75 (1.89): nothing to fix. Raise budget in steps and keep creative volume up.
Moving from the median to the top quarter is mostly offer and creative work, which is what the Icy Bear case study shows (£30k to £500k a month in 8 months) and what our Meta ads team does for clients. If you report blended revenue, keep it in MER against your own history; none of these percentiles are blended.
If you are comparing agencies, our guide to Meta ads agencies for DTC brands sets out the options and who each one suits, and our guide to DTC ad creative agencies does the same for creative partners.
FAQ
What is a realistic CPA for a DTC brand on Meta ads?
Across 79 DTC brands in the 90 days to 28 October 2026, the median Meta CPA was $63.80 per purchase, with the middle half between $41.76 and $98.54; 11.4% paid $30 or less and 74.7% paid $100 or less. A realistic target sits inside that range and below your break-even CPA.
Is a 3x ROAS on Meta ads realistic?
Rarely, on first-order Meta-attributed revenue. Over the 90 days to 28 October 2026, 7.6% of 79 DTC brands had a ROAS of 3.0 or above; over 12 months, 7.3% of 82 brands did. The 90th percentiles were 2.38 and 2.47, so 3x is a top-decile result in this sample.
What is a good click-through rate and conversion rate for Meta ads?
The median link CTR was 1.70% and the median conversion rate (purchases divided by link clicks) 2.68% across 57 DTC brands between 31 July and 7 September 2026. The middle half sat between 1.21% and 2.42% on CTR and 1.54% and 3.72% on conversion rate; above 2.42% and 3.72% is the top quarter on each.
Does a ROAS under 1 on Meta mean I am losing money?
A first-order Meta-attributed ROAS under 1 means the first order returned less revenue than the ads cost. Over the 90 days to 28 October 2026, 69.6% of 79 DTC brands were at or above 1.0, so about three in ten sat below it. Strong repeat purchase can still make that profitable on lifetime value, which this dataset does not measure.
Sources
- MHI Meta Benchmarks, DTC, October 2026 edition, MHI Media; hub post with methodology, 22 October 2026
- MHI Media warehouse of client Meta ad accounts, benchmark cuts edition, read-only extract 29 October 2026, data settled to 28 October 2026