Meta Ads Average ROAS in 2026: Why the Number You're Chasing Is a Trap

Every week an owner asks me some version of the same question: "What ROAS should I be getting on Meta?" And every week I have to disappoint them, because the honest answer is that the question is built on a trap. There is no single average ROAS worth chasing, and the businesses fixated on hitting some benchmark number are usually the ones measuring the wrong thing entirely.

Let me explain why, give you real ranges anyway because you deserve something concrete, and then show you the number that actually matters — the one tied to money in your bank account rather than a figure Meta invents for you.

Why "average ROAS" is close to meaningless

Start with what ROAS even is: revenue divided by ad spend. Spend a thousand dollars, generate four thousand in revenue, that is a 4x ROAS. Simple. The problem is that every word in that sentence is slippery.

Which revenue? Blended or new-customer only? Attributed how, and over what window? Reported by whom — Meta, or your actual books? Two businesses can both say "we run at 4x" and mean completely different things, one of them profitable and one quietly going broke.

Here is the trap in one sentence: a "good average ROAS" for a med spa selling a two-hundred-dollar facial and a roofer selling a twenty-thousand-dollar re-roof are not remotely the same number, and comparing yourself to a blended benchmark across industries tells you nothing.

Chasing someone else's average ROAS is like asking what a good salary is without saying the job, the city, or the hours.

The blended-versus-new-customer trap

This is the one that fools the most owners, so let me be blunt about it. Blended ROAS counts all revenue, including repeat customers and referrals who would have bought anyway. New-customer ROAS counts only revenue from people the ads actually acquired.

Blended always looks better. That is exactly why it is dangerous. If half your revenue comes from repeat customers, your blended ROAS can read a healthy 5x while your ads are barely breaking even on the new business they are genuinely responsible for. You feel great and scale up a campaign that is actually treading water.

For local service businesses, new-customer ROAS is almost always the number you should optimize on, because the entire job of Meta is acquisition. Your existing customers are not seeing your prospecting ads and deciding to come back; they are coming back because you did good work. Do not let their loyalty flatter your ad account.

Attribution windows quietly change your number

Meta lets you attribute conversions across different windows — a one-day click, a seven-day click, longer view-through settings. Widen the window and your reported ROAS goes up, because you are crediting the ad for more of what happens after the click. Narrow it and the same campaign looks worse.

Nothing about the actual business changed. Only the accounting did. This is why comparing your ROAS to someone else's is often meaningless — you may be measuring a seven-day click window against their view-through-inclusive number and drawing conclusions from an apples-to-oranges gap. Pick a window, hold it constant, and judge trends against your own history rather than against strangers on the internet.

The big one: platform-reported ROAS lies without offline tracking

Now the failure that costs local businesses the most money. Meta reports ROAS based on the conversions it can see — usually a form submission, a click-to-call, a lead. But for a service business, a lead is not revenue. A lead is a maybe. The revenue happens later, offline, when someone books, shows up, and pays.

So Meta shows you a lovely "4x ROAS" based on lead value it assigned, while your actual booked-and-paid ROAS might be 2x or 8x — and you have no way of knowing which, because Meta never sees the sale. It is optimizing toward cheap leads, not toward leads that turn into paying customers. Those are wildly different things, and the gap between them is where budgets get wasted.

The fix is offline conversion tracking: feeding real booked-job data back into Meta so the platform learns which leads actually became revenue and optimizes toward the people who look like your paying customers, not the people who look like form-fillers. Without it, you are flying on instruments that were never connected to the engine.

Honest benchmark ranges, by situation

You still want numbers, so here they are — with the loud caveat that these are new-customer ROAS ranges for local service businesses, and your mileage will vary with market, offer, and how you count.

Notice that a "good" number ranges from 2x to well over 9x depending entirely on ticket size and repeat value. That is precisely why chasing a universal average is a fool's errand.

The levers that actually move ROAS

If the number varies this much, what do you actually pull to improve it? Four things, in rough order of impact.

Measure it against real booked revenue

Here is where all of this lands. Stop asking Meta what your ROAS is. Meta is measuring leads it can see, on an attribution window you may not have chosen, blended with revenue it should not be counting.

Instead, build the loop that measures booked, paid revenue against ad spend: track every lead to its outcome in your CRM, feed the closed jobs back to Meta as offline conversions, and calculate cost per booked job. That single number — what you paid in ads to produce one paying customer — cuts through every accounting illusion above. It cannot be flattered by repeat customers, widened windows, or lead-value guesses. It is just what a customer costs versus what a customer is worth.

Once you are measuring that, the "average ROAS" question dissolves. You stop comparing yourself to strangers and start asking the only question that matters: does one more dollar of spend produce more than a dollar of booked profit? If yes, spend more. If no, fix creative, offer, page, or speed until it does.

If you want that loop built — real offline tracking, ads and instant follow-up wired together, ROAS measured against jobs you actually banked — that is the work. Book a free strategy call and we will show you what your real number is, not the one Meta wants you to see.