ROAS 5 but no money: why a “good report” ≠ profit

TL;DR

A client arrives with a Meta Ads report: ROAS 5.2, spent $3,000, revenue $15,600. Looks perfect. But two months later the owner says: “there’s no money in the account, I’m in the red.” How? Because ROAS counts revenue, not profit. If the cost of goods is 70%, taxes 20%, logistics 15%, then of $15,600 in revenue what’s left after all expenses is a minus. In this article we break down a real failure case, 4 root causes why a “good” metric can mask losses, and how to diagnose the economics BEFORE launching ads. At LeadPrice we’ve seen this picture in 6 out of 10 new clients who came after other agencies — the report is pretty, the business is burning.

A failure case: a clothing e-commerce store with a “successful” ROAS of 5

March 2023. The owner of an online clothing store (name withheld under NDA) approached us. The previous agency had run Meta Ads for 4 months, reporting a ROAS of 4.8-5.3 every month. The client was satisfied — until the moment their accountant showed them the real P&L.

What the agency’s reports showed (over 4 months):

• Ad spend: $11,200
• Revenue (attribution window 7 days click): $56,000
• ROAS: 5.0
• Number of purchases: 1,120 (average ticket $50)

What the real accounting showed:

• Cost of goods (purchased from Turkey): 68% of price = $38,080
• Logistics (international + domestic delivery): $4,200
• Product returns (20% of orders): $11,200 in losses
• Payment system fee (3.5%): $1,960
• Taxes (simplified 5% + social contributions): $3,360
• Agency fee (15% of ad spend): $1,680
• Ad spend: $11,200

The total: Revenue $56,000 − cost of goods $38,080 − logistics $4,200 − returns $11,200 − fees $1,960 − taxes $3,360 − agency $1,680 − ads $11,200 = minus $14,680 over 4 months.

The agency showed a ROAS of 5, while the business was losing $3,670/mo. The report was technically correct — Meta really did count $56K of revenue by attribution. But nobody checked the unit economics BEFORE scaling. The previous agency worked in the model “pour in traffic → improve ROAS → report to the client.” The real math stayed off-screen.

Root cause #1: ROAS counts revenue, not margin

ROAS (Return on Ad Spend) = revenue / ad spend. The formula takes revenue from Meta Events Manager or Google Analytics, but doesn’t account for cost of goods, logistics, returns, taxes. If your gross margin is 25% (cost of goods + logistics take 75%), then a ROAS of 5 means not profit but possible losses after all expenses.

An example of the math for e-commerce:

A ROAS of 5 looks like success. But if the gross margin is thin, the business runs at a loss. The right metric for e-commerce isn’t ROAS but MER (Marketing Efficiency Ratio) = total revenue / total marketing spend, or better still POAS (Profit on Ad Spend) = net profit / ad spend. If POAS is negative, the advertising is burning money, even if ROAS is 10.

At LeadPrice on the first call we ask: what’s your gross margin? If the client doesn’t know, that’s a red flag. Ordinary agencies at this stage say “OK, let’s launch a $500 test budget.” We don’t. First we calculate the Target ROAS based on the real economics, then validate whether we can reach it in the auction. If not, we say honestly: “your model won’t sustain paid advertising at current prices.”

Root cause #2: The attribution window ≠ real customer behavior

By default Meta counts conversions in a 7 days click + 1 day view window. That is, if a user clicked the ad and then bought within 7 days (even via Google search or a direct visit), Meta credits itself. If they saw the ad (but didn’t click) and then bought within 1 day, it also credits itself.

The problem: the real customer journey in B2C is often 14-30 days, in B2B 30-90 days. A user saw the ad → went to google reviews → compared with competitors → came back 2 weeks later via organic search and bought. Meta will NOT count that purchase in its ROAS (because the attribution window has ended), but in reality the ad worked as the first touchpoint.

The reverse situation: a user accidentally clicked the ad (a mis-click on mobile), came back 3 days later via Google Ads and bought. Meta will credit the purchase to itself in the report, though the real conversion channel is Google. The result: Meta’s ROAS is inflated, Google’s ROAS is underestimated, and the business owner doesn’t understand the real picture.

What it looks like in real numbers (the case of an aesthetic medicine clinic we manage):

• Meta Ads report (7 days click): 120 consultation bookings a month, ROAS 4.2
• Google Ads report: 80 bookings, ROAS 3.1
• CRM (real data): 180 bookings in total, of which 15% said “I saw the ad on Facebook” (but came via Google or directly)

The sum of conversions in the Meta + Google reports = 200; in reality in the CRM = 180. Overlap (duplicated attribution) = 20 bookings, or 11%. If you count ROAS separately by channel, everything looks great. If you look at all of marketing together (MER), the picture is different.

The solution: end-to-end analytics with UTM tags + CRM integration. At LeadPrice we build a single dashboard where we see First Touch (the first channel of acquaintance), Last Touch (the last channel before conversion), and the whole path in between. It isn’t perfect attribution (there’s no such thing), but it gives the real picture rather than the rose-tinted glasses of Meta Events Manager.

Root cause #3: Manipulating the attribution model in reports

Some agencies (not all, but many) play with the attribution settings so the report looks better. For example:

• Switching on 28 days click attribution instead of 7 days — ROAS inflates 15-30%, because Meta credits conversions that actually came through other channels
• Including Engaged view attribution (the user watched a video for 10+ seconds, then bought within 1 day) — ROAS grows another 10-20%, but the real reason for the purchase could have been completely different
• Not segmenting New vs. Returning customers — retargeting to a warm base gives a ROAS of 8-12, cold traffic a ROAS of 2-3, and in the overall report an average ROAS of 5 looks decent. But if 80% of the budget goes to retargeting, you aren’t scaling the business, just bringing the same people back

An example from practice (an e-commerce client who came to us after another agency):

The previous agency reported a ROAS of 6.3 at a 28 days click attribution window. We switched to 7 days click (Meta’s standard for e-commerce) — ROAS fell to 4.1. Then we split New vs. Returning:

• Retargeting (20% of budget): ROAS 11.2
• Cold Lookalike traffic (50% of budget): ROAS 2.8
• Cold Broad traffic (30% of budget): ROAS 1.9

Weighted average ROAS = (11.2 × 0.2) + (2.8 × 0.5) + (1.9 × 0.3) = 2.24 + 1.4 + 0.57 = 4.21. That’s the real number. But if you show the client only “overall ROAS 6.3 over 28 days,” they think everything’s fine until they see the bank account.

Our standard: we report on 7 days click (Meta) or Last non-direct click (Google), segment New vs. Returning, show MER separately. If a client asks “why is your ROAS lower than the previous agency’s?”, we explain the difference in methodology. We don’t play “pretty numbers,” we play “real profit.”

Root cause #4: Ignoring LTV and CAC over the long term

ROAS is a short-term metric. It shows how much revenue you got from ads now. But it doesn’t show:

• How much this client will bring over 6-12 months (LTV — Lifetime Value)
• How much it costs to acquire a client across all marketing channels (CAC — Customer Acquisition Cost)
• Whether this client will pay back through repeat purchases

Example: a SaaS product with a $50/mo subscription. A client buys after seeing an ad; the first month brings $50 of revenue, ad spend $30 → ROAS 1.67 (seemingly poor). But if the average subscription length is 8 months → LTV = $400. CAC $30 → LTV/CAC = 13.3 (very good). Looking only at the first month’s ROAS, it seems the ads don’t work. Counting LTV, the business is in the black.

The reverse example: e-commerce with impulse purchases. Average ticket $60, ROAS 5 ($10 of ad spend brings $50 of revenue per purchase). Repeat purchase rate 5% (only 5 of 100 clients buy a second time). LTV = $60 × 1.05 = $63. CAC = $10 in ad spend, plus the agency fee, creatives, etc. — realistically CAC is $15. Gross margin 30% → net profit per client $63 × 0.3 = $18.90. CAC $15 → $3.90 of profit per client. It works, but the margins are thin. If CAC rises to $20 (CPM got more expensive), the business is in the red, even if ROAS is still 5.

In our methodology (the LeadPrice approach to the funnel) we always calculate the LTV/CAC ratio BEFORE launching ads. If LTV/CAC < 3 → the model doesn’t scale without optimization (you need to either lower CAC through creatives/targeting or raise LTV through retention/upsell). If an agency doesn’t do this, it’s just pouring in traffic and praying the economics work out by themselves.

How to diagnose the real economics of advertising: a pre-launch checklist

Before pouring budget into Meta/Google/TikTok, go through this checklist. If the answer to even 2 questions is “I don’t know,” the advertising will burn money with 80% probability, even if ROAS looks good.

1. What’s your Gross Margin?

Formula: (Revenue − Cost of goods − Logistics) / Revenue × 100%. If it’s under 40% for e-commerce or under 60% for services, paid advertising will be hard. Target ROAS must be at least 1 / Gross Margin. So if Gross Margin is 30% → Target ROAS > 3.33 to break even after ad spend.

2. What’s your Average Order Value (AOV)?

If the average ticket is $20 and the CPA (cost per acquisition) in your niche is $15 → the margins are thin. If AOV is $200 and CPA is $15 → there’s room to maneuver. A low AOV doesn’t mean “don’t run ads,” but it does mean “you need to work on upsell/cross-sell in the funnel.”

3. What’s your Repeat Purchase Rate?

What % of clients buy a second time within 6 months? If 0-5% (one-time purchases) → LTV equals AOV, and that has to be built into the Target CPA. If 30-50% → LTV is 1.5-2 times higher, you can afford a higher CPA on the first order.

4. What’s your current Conversion Rate on the site?

If CR < 1% for e-commerce or < 3% for lead-gen → the problem isn’t the traffic but the funnel (site, offer, price, trust). Launching ads onto a weak funnel = burning the budget. First fix conversion, then scale traffic.

5. Do you have end-to-end analytics (CRM + GA4 + ad platforms)?

If not, you won’t see the real picture. Meta says 100 purchases, Google says 80, the CRM says 120 — who’s right? Without a Single Source of Truth you’ll be making decisions based on guesswork.

6. What’s your Target CAC to be profitable?

Formula: Target CAC = LTV × Target Profit Margin. If LTV is $100 and you want a 30% profit margin → Target CAC = $100 × 0.3 = $30. Now the question: is a $30 CPA realistically achievable in your niche’s auction? If the average CPA in the niche is $50 → you need to either raise LTV (retention, upsell) or lower CAC (creatives, landing page).

At LeadPrice we do this diagnostic at the first meeting. If we see the economics don’t work, we say so honestly. In 2 out of 10 cases we turn the client down, because we understand: under current conditions advertising won’t bring profit, even if ROAS is 10. Ordinary agencies at this stage say “let’s try” — because they need to hit their sales plan. We work for the long term, so we filter at the entrance.

What to do if you have a ROAS of 5 right now but no money

Step 1: Calculate the real profit per unit sold

Take one average order and write out all the costs: cost of goods, logistics, packaging, payment fee, returns (the average % in your niche), taxes, staff salaries (proportionally). What’s left after that is the Gross Profit per order. If it’s less than CAC, you’re in the red, regardless of ROAS.

Step 2: Check the attribution settings in Meta/Google

Meta Ads Manager → “Columns” tab → “Customize Columns” → “Attribution Setting.” It should be 7-day click for e-commerce, 1-day click for high-ticket or long-cycle (so as not to inflate the numbers). If it’s set to 28-day click or Engaged view, ROAS is inflated 20-40%.

Step 3: Segment New vs. Returning customers

Create separate ad sets for cold traffic and retargeting. Look at ROAS separately. If all the profit comes from retargeting, you aren’t scaling, you’re circling the same audience. Cold traffic needs fixing (creatives, offer, landing page).

Step 4: Make MER the main metric

MER (Marketing Efficiency Ratio) = Total Revenue / Total Marketing Spend. It includes ALL channels (Meta, Google, influencers, email, SMM — everything you spent on). If MER < the Target ROAS (which you calculated in Step 1) → marketing as a whole is ineffective, even if one channel shows a ROAS of 5.

Step 5: Audit the funnel

Maybe the problem isn’t the ads but conversion. If CR < 1.5% (e-commerce) or < 5% (lead-gen) → the site/landing page needs fixing first. Raising CR from 1% to 2% = a 2x cheaper CAC on the same traffic. That’s often more effective than trying to squeeze +10% ROAS out of the ads.

Red flags: when an agency shows a “good ROAS” but something’s off

• The agency doesn’t ask about your Gross Margin at the first meeting — if the strategist doesn’t know the economics, they’re steering blind
• The reports show only ROAS, no CAC or MER — they may be hiding the real picture
• An attribution window over 7 days for e-commerce — they’re inflating the numbers
• They don’t segment New vs. Returning — all the ROAS may be coming from retargeting while cold traffic is in the red
• They promise a “guaranteed ROAS of 5+” — nobody controls the auction; guaranteeing ROAS is impossible
• They don’t integrate the CRM with the ad platforms — they work from Meta/Google reports, not from your business’s real data

If you see 2+ points from this list, it’s worth reviewing the cooperation. At LeadPrice we accept ~20% of incoming requests, because we deliberately filter out clients we won’t be able to deliver results for. Our reputation is worth more than one contract. So if we see the economics don’t work, we say so openly. Sometimes we recommend first working on the product/pricing, then returning to advertising. That’s honest.

FAQ: the most common questions about ROAS and real profit

What ROAS is considered good for e-commerce?

There’s no universal answer. It depends on Gross Margin. If Gross Margin is 30% → Target ROAS must be at least 3.33 to break even (not counting other expenses). If Gross Margin is 60% → Target ROAS 1.67. Plus you have to factor in the agency fee (10-20%), taxes, operating costs. In practice, for clothing/electronics e-commerce a healthy ROAS = 4-6. For premium goods with a high margin = 2-3. For dropshipping with a 20% margin = 6-8 (otherwise you won’t survive).

Why is ROAS on Meta higher than on Google?

Two factors. First: Meta counts conversions in a wider attribution window (7 days click + 1 day view), while Google defaults to Last non-direct click (the last click before conversion). That is, Meta credits conversions that may have come through other channels. Second: the audience. Meta works well on warm/hot traffic (retargeting, lookalikes on buyers), so ROAS is higher. Google Search catches demand (the user is already searching) → more often cold traffic → lower ROAS but higher volume. Both channels are needed; comparing them head-on is incorrect.

The agency says the ads need 3 months to “mature.” Is that true?

Partly true. Meta’s Learning Phase algorithm really does need ~50 conversions per ad set to stabilize (that’s 1-4 weeks depending on budget). But if after 3 months ROAS is still negative and the agency says “a little longer,” that’s a red flag. In normal practice the first signals (is the hypothesis working) are visible in 2-4 weeks. If CPA is 2 times above Target, you need to change the creatives/targeting/offer, not wait “until the algorithm learns.” The algorithm won’t work a miracle if the economics don’t add up.

How do I know whether the problem is the ads or the product?

Look at the Conversion Rate. If CR on the landing page is < 1% for e-commerce, the problem is the funnel (a weak offer, a high price, no trust, a complicated checkout). If CR is 2-3%, the funnel is OK, you can scale traffic. If CR > 3% but CAC is still high, the problem is targeting or creatives (you’re showing to the wrong audience, or the creative doesn’t hook). The typical mistake: pouring more budget onto a weak funnel hoping “it’ll somehow work.” Spoiler: it won’t.

Which is better: high conversion (low CPA) or high AOV?

It depends on the business model. For subscription or high-LTV products a high AOV is better, even if CPA is higher (because it pays back through repeats). For impulse purchases (low LTV) a low CPA is critical, otherwise you won’t reach profit. Ideally — balance. In practice we often see: a business tries to raise AOV through upsell, but CR falls (because the user isn’t ready to buy the more expensive item). You have to test.

Can a ROAS of 10+ be achieved consistently?

Yes, but only on retargeting or in specific niches (for example, digital products with a 90% margin). On cold traffic in competitive niches a ROAS of 10 is an outlier, not the norm. If an agency promises a “consistent ROAS of 10 on cold traffic,” they’re either working in an unrepresentative niche or misleading you. In our experience (80+ cases on the site) a healthy ROAS on cold traffic = 3-6 depending on the niche. If you’ve achieved more — great, but don’t build it into the forecast.

Conclusion: a metric isn’t a result — the result is money in the account

A ROAS of 5 is a pretty number in a report. But if after all expenses (cost of goods, logistics, returns, taxes, agency fee) you’re left with a minus, the business is burning, even if the agency says “look what a great ROAS.” The real metric of success is net profit per client (unit economics) and the LTV/CAC ratio. If they’re healthy, the business scales. If not, any ROAS is an illusion.

At LeadPrice we don’t work on the “pour in traffic → report ROAS” model. We work as a business partner: first we diagnose the economics, then build the funnel around the real numbers, then scale only what’s proven to work. It’s slower than “we launched a campaign in 3 days,” but it’s honest. If you want an agency that tells the truth (even when it’s unpleasant) rather than selling a “guaranteed ROAS of 10” — write to us. We’ll look at your numbers and say whether we can help or not. In 8 out of 10 cases we can. In 2 out of 10 we say honestly: “your model isn’t ready for paid advertising yet; here’s what needs fixing first.”

Because our reputation is built not on the number of contracts signed, but on the number of businesses that actually made money after working with us. And that’s the only metric that matters.