What is a good ROAS for Facebook ads?
A good ROAS is anything meaningfully above your break-even ROAS, and break-even ROAS equals one divided by your gross margin. At 50% margin you break even at 2.0, at 30% margin at roughly 3.3, at 70% margin at about 1.4. There is no universal good number, because the same 2.5 ROAS is profitable for one business and loss-making for another, purely as a function of margin.
Last updated 2026-08-11
The break-even formula
ROAS is revenue divided by ad spend, so the break-even point is where the margin on that revenue exactly covers the spend. Algebraically that is one divided by gross margin expressed as a decimal. A store keeping 40 cents of each revenue dollar after product, shipping and processing costs needs $2.50 of revenue per ad dollar just to break even, hence 2.5. Compute your margin honestly, including fulfilment, payment fees and packaging, because every cost you forget flatters the break-even and lets unprofitable campaigns look fine. This one division is the anchor for every ROAS judgment in the account; without it the metric is a number without a meaning.
From break-even to a working target
Break-even ROAS covers cost of goods and the ad spend, and nothing else: no salaries, no software, no profit. A working target sits above it by whatever contribution you require. One clean way to set it: decide the profit you want per revenue dollar, subtract it from margin, and recompute. If margin is 50% and you want 15 points of profit, you are effectively working with 35%, giving a target near 2.9 rather than 2.0. Businesses with strong repeat purchase can justify running closer to first-order break-even because later orders carry the profit, but that is a deliberate payback decision, not a default.
High ROAS is not automatically good news
ROAS is a ratio, and ratios improve when the denominator shrinks. An account showing a very high ROAS on small spend is often just harvesting its cheapest demand, retargeting and brand-adjacent audiences, while leaving profitable growth unbought. If ROAS sits far above target, the interesting question is whether spend should rise, accepting a lower ratio on the incremental dollars, because the goal is maximum total contribution, not maximum ratio. The general shape is that marginal ROAS declines as spend scales into colder audiences, so businesses that manage to the ratio alone systematically under-invest. Manage to profit at a spend level, with ROAS as the instrument.
Attribution decides what the number even is
Reported ROAS depends on the attribution settings behind it. A window including view-through credit reports a higher ROAS than click-only; Meta's number and your analytics' number will disagree because they model credit differently; and platform-reported revenue is not the same as incremental revenue, since some attributed buyers would have purchased anyway. This does not make the metric useless, it makes consistency mandatory: pick one attribution basis, compute break-even against the same basis, and never compare a ROAS from one window to a target derived from another. When the reported number and the bank account disagree over a quarter, the bank account is right.
Using ROAS and CPA together
ROAS and CPA are the same economics viewed from different ends, revenue per dollar versus cost per order, and each has a blind spot the other covers. ROAS quietly rewards campaigns that skew toward high-value orders even when volume collapses, while CPA misses order-value differences entirely. Reading both catches what one alone hides: a rising ROAS with falling order volume is skew, not improvement, and a stable CPA on shrinking AOV is quiet erosion. For accounts with wide product price ranges, ROAS is usually the primary lens with CPA as sanity check; for single-product accounts they are interchangeable and CPA is simpler to reason about.
Judging ROAS on enough data
ROAS on a handful of conversions is noise, and a single large order can double a day's figure without meaning anything. Judge on windows long enough to hold a reasonable event count, compare like periods to respect weekly and promotional cycles, and be suspicious of any conclusion that reverses when you shift the window a few days. During learning phases, expect the number to be unstable and worse than steady state. The practical cadence for most accounts is directional daily glances but decisions on weekly or multi-week windows, sized so that the decision would survive one lucky or unlucky order being removed.
Every figure on this page is arithmetic on your own margin, not an industry benchmark. Published average ROAS numbers blend attribution settings, verticals and margin structures, which is why chasing someone else's ratio is a category error.