ChatGPT
Claude
Gemini
Grok
Perplexity
Your Facebook ads are working only if they bring in new customers at a profit you can prove outside the ad platform. Meta can show a green "active" status and a strong return on ad spend, and your business can still lose money on every order. So the real test is not a number inside Ads Manager. It is your own math: what you can afford to pay for a customer, your monthly profit and loss, and how much profit your new customers bring in.
The Platform's "Working" Is Not Your "Working"
Meta's dashboard is built to show you Meta's version of success. A campaign can spend its full budget, report a healthy ROAS (return on ad spend, your sales divided by your ad spend), and still leave your bank balance flat.
That gap has two parts, and you have to separate them:
- The reported number might not be trustworthy. If the tracking is off, every judgment built on it is off too.
- Even a true number does not tell you if you made money. Sales is not profit, and ad-account sales are not the same as money in the bank.
The rest of this guide fixes both.
First, Make Sure the Number Is Even Trustworthy
Before you judge anything, do one quick check: compare the revenue Meta reports to your Shopify sales for the same dates. If they are far apart, your tracking needs cleaning up before any decision, and that is its own topic. (See our guide on why your ROAS does not match your bank account.)
There is also a timing trap from earlier this year.
A moving number is not the same as moving performance.
What "Working" Really Means: Profit You Can Prove
Once you trust the tracking, judging ads comes down to two layers, in order. Work them top to bottom: profit first, then effectiveness.
Layer 1 · Profitability
Your unit economics set what you can afford to pay for a customer, and your profit and loss statement proves whether you actually did.
Layer 2 · Effectiveness
Your new-customer contribution margin shows whether the ads are creating growth or just riding on customers you already had.
Layer 1, Part One: Know What You Can Afford to Pay for a Customer
Start with your unit economics: every cost of delivering one order. Add up your product cost (COGS), shipping, fulfillment, transaction fees, and returns. That total is your cost of delivery. What is left over is what you can spend to get a customer, which is your CAC (cost to acquire a customer). Here is the math on a $50 product, a simple filter.
| Shipping | $5.00 |
| Fulfillment (3PL) | $2.00 |
| Product cost (COGS) | $10.00 |
| Transaction fees | $1.80 |
| Returns (at a 7% return rate) | $3.50 |
| Cost of delivery | $22.30 |
| Product price | $50.00 |
| Minus cost of delivery | −$22.30 |
| Breakeven CAC | $27.70 |
| Minus target profit (20% margin on $50) | −$10.00 |
| Target CAC (the number to hold ads against) | $17.70 |
One more reality check: your target CAC has to be reachable at real market ad rates. If it costs more than $17.70 to get a customer in your market, the ads are not the problem, the offer is. Your levers are price, offer, and bundling. Raise the order value or the margin until the target CAC is one you can actually hit.
Layer 1, Part Two: Your Profit and Loss Is the Real Test, Not the Platform
The ad account is not your profit and loss statement. A proper P&L is the profitability test, and the best operators check it monthly at least, and daily when spend is high. Two things a good P&L does that the platform cannot: it separates variable costs from fixed costs, and it shows the truth when the platform looks fine. That second one is where ads quietly go wrong.
| In the ad account | In the P&L |
|---|---|
| Sales looked steady | Losing money on every new customer |
| ROAS looked fine | The CAC math did not clear breakeven |
| Shopify payouts kept arriving | Daily P&L tracking caught the loss |
The ads "worked" by every number Meta showed. The P&L told the real story. The fix was not a new campaign. It was restructuring the front-end offer and adding checkout upsells to lift the average order value back into profit.
And keep this in mind as you grow: the larger you scale, the less efficient you usually get. A CAC you could afford at $5,000 a month may not hold at $50,000. Affordability is not a one-time calculation. Re-check it as spend climbs.
Layer 2: New-Customer Profit Shows If the Ads Are Doing the Work
Profitable overall is good. But you still need to know the ads are creating growth, not taking credit for customers you already had. The tool for this is contribution margin (what is left from a sale after you subtract the variable costs). Run it two ways:
- Split new customers from repeat buyers. Pull the new-versus-existing customer data from Shopify.
- Count all your ad spend against new customers only. This is a strict, honest read. If the ads still clear a profit after that, they are earning their place.
Your new-customer contribution margin is the effectiveness signal. If it is healthy, the ads are buying real growth. If most of your ad-driven "sales" are repeat buyers who would have come back on their own, the ads are riding on demand you already had. How hard you work this depends on your spend:
Under $20K/month per channel
Keep it simple. Run one campaign, do not over-segment, and read new versus existing in Shopify. Splitting too early starves the campaign of the data it needs to learn.
Over $20K/month per channel
Segment new versus existing with separate audiences and exclusions, so the numbers get sharper and you can steer spend toward acquisition.
Platform "Incrementality" Tools Are Context, Not the Verdict
You will see tools that promise to measure the ads' true impact. Use them for context, never as the final word.
Meta offers Incremental Attribution (a holdout-style model that tries to strip out purchases the ad did not actually cause). It is best on new-customer acquisition, and it is worth treating as an experiment, not a default. Third-party tools like Triple Whale and Northbeam can help too, and both offer deterministic view-through measurement.
Here is why none of them replace your P&L: platforms tend to overstate their own impact, and independent cross-checks routinely confirm it. Let these tools inform your read. Judge with your own numbers.
Give the Ads Enough Time and Data to Judge Honestly
Do not crown or kill a campaign on a few days of numbers. Ads need a learning period and enough conversion volume before the data means anything. Judging too early is how good campaigns get shut off and bad ones get scaled. (We cover how long to wait in a separate guide.)
Why Your Agency Shows You the Platform Number
If your agency reports ROAS and calls it a day, that is usually the incentive, not bad intent. An agency is hired to manage a channel, so it reports the channel's metric. The business-level math, your P&L and your new-customer profit, sits outside what most agencies are asked to own.
So own it yourself. Set the business KPI, then ask for it: your target CAC, and your contribution profit on new customers with ad spend counted against them. A good partner will welcome the question. If asking for the real number makes the reporting fall apart, that is your answer.
Do This Now to Judge Your Own Facebook Ads
- 01
Pull your unit economics.
List every cost of delivering one order, then total your cost of delivery.
- 02
Compute breakeven and target CAC.
Price minus cost of delivery is breakeven. Take out your target margin to get the CAC to aim for.
- 03
Build a P&L.
Monthly at least, daily when spend is high, with variable and fixed costs separated.
- 04
Pull new versus existing customers from Shopify.
For last month, and for the current month so far.
- 05
Split contribution margin, spend on new only.
If new-customer contribution profit is positive with all ad spend counted against it, the ads are working. If not, fix the offer before you spend more.
Want to run your own numbers with help? Paste this into any AI tool:
AI Prompt
Copy this and paste it into ChatGPT or Claude. It will walk you through your own numbers.
You are helping me judge whether my Facebook ads are actually profitable. Here are my numbers: - Product price: $___ - Cost of delivery per order (COGS, shipping, fulfillment, transaction fees, returns): $___ - Current cost to acquire a customer (CAC): $___ - New vs returning customer split last month: ___% new - Ad spend last month: $___ - New-customer revenue last month (from Shopify): $___ Walk me through, one step at a time: my breakeven CAC, my target CAC at a 20% margin, and whether my new-customer contribution margin is positive when I count all ad spend against new customers only. Point out where my ads look like they are working but are not.
See more profit-first playbooks in our insights library.
Common Questions
Facebook Ads Profitability FAQ
How do I know if my Facebook ads are actually profitable?
Do not answer this inside Ads Manager; check three things in your own numbers. First, know what one order really costs you to deliver, and what that means you can afford to pay for a customer. Second, check your profit and loss statement monthly, because ads can look fine while the business loses money. Third, look at profit from new customers only, with all your ad spend counted against them. If new customers are profitable after ad spend, your ads are working.
Why did my Meta conversions drop in March 2026?
Most likely nothing is wrong with your ads. On March 3, 2026, Meta changed what counts as a click: only real link clicks count now, and likes, shares, and saves moved to a separate category. Many accounts saw reported conversions drop with no real change in sales. Do not compare your numbers from before and after March 2026 as if they mean the same thing. Re-baseline your reporting from mid-March forward.
Should I trust Meta's ROAS number?
Trust it for what it is, not for what it is not. ROAS tells you what Meta tracked, and after you verify the tracking against your Shopify sales, it is useful for comparing one campaign against another. What it cannot tell you is whether you made money. Sales are not profit, and Meta's number does not know your costs. Judge campaigns inside the platform if you want, but judge your business on your own profit and loss statement and your new-customer profit.
Should I use Meta's Incremental Attribution tool?
Yes, Meta's Incremental Attribution tool is worth testing if you are a direct-to-consumer ecommerce brand with reliable website purchase tracking. It helps Meta optimize toward sales that likely would not have happened without the ad. Results are less clear for brands with significant Amazon or retail sales, since Meta cannot see every purchase. Start with one established campaign and judge the results using total revenue, new-customer acquisition cost, and new-customer contribution margin, not Meta ROAS alone.
The Bottom Line on Whether Your Facebook Ads Are Working
Meta's "delivering" is not the same as "working." Ads are working when new customers come in at a profit you can prove in your own P&L, not when the dashboard looks green. Get your unit economics, hold ads against a real target CAC, and read your new-customer contribution margin. That is a test the platform cannot fake and an agency cannot spin.
