Ads Manager can show 4.2x ROAS while your business clears 1.6x after costs. If you spend $3k to $30k a month on Meta and sell something with a cost of goods, my pick is a flat-fee done-for-you service: one whose work you can audit against revenue that actually cleared, not just the number Meta attributes to itself.

I managed Meta and Google accounts by hand for most of a decade. Before I look at a case study now, I ask one question: Which ROAS are you answering to? Everyone says they improve it. Almost no one means the same number.

Three numbers hide under that word:

  • Platform ROAS is Meta-attributed revenue divided by Meta spend. It tells you what Ads Manager claims.
  • Blended ROAS, often called MER, is total revenue divided by total marketing spend. It tells you what happened across the business, though not which channel deserves the credit.
  • Margin ROAS brings product, shipping and fees into the calculation. It asks whether the sales left enough contribution to pay for the advertising.

The break-even math is brutal. At 50% margin, you need 2.0x revenue over ad spend just to stand still; at 40%, you need 2.5x; at 20%, you need 5x. Management fees raise those hurdles. Attribution can widen the gap between the report and the business, too: with 7-day click and 1-day view windows, platforms can claim the same sale. That gap is where bad hiring decisions live.

The test behind this ranking

I rank providers by the number they let you fire them over. Not the number in the pitch deck. The number you can check against money that cleared. I look for four things:

  1. What does the fee reward?
  2. Which ROAS gets reported when the numbers disagree?
  3. Who makes and tests the creative?
  4. What can you audit without asking permission?

If a provider controls the tracking and bills on that tracking, I mark it down. If the fee rises simply because spend rises, I mark it down. Follow the incentive and you can usually predict the report.

A fee changes break-even before a single ad runs. Say you spend $1,500 on ads at 50% margin. Without a management fee, you need $3,000 in revenue to break even, or 2.0x. Add a $750 fee and you need $4,500, or 3.0x, using (ad spend + fee) / (ad spend × margin) as the break-even calculation described here. Most sales calls never show you that second number. I used to quote the first and hope no one asked about the second. I was wrong.

Rank and providerFee rewardsNumber to challengeCreativeWhat you can audit
6. AI toolsSoftware tierPlatform ROASYou make itRules and alerts; no owner
5. Percent-of-spend agencyHigher spendPlatform vs blended ROASTeam volumePlatform report and your own sales
4. Pay-per-result shopBillable eventsBilled metric vs qualified pipelineVariesInvoice against your CRM
3. FreelancerRetained workBlended ROASOne person’s capacityDirect account access
2. In-house hireFixed salaryMarginIn-house paceYour P&L
1. Flat-fee done-for-youFixed feeBlended and marginOngoing testingAction log against your numbers

6. AI optimization tools: useful automation, no one to fire

For: the owner spending under about $3k a month who wants automation without hiring anyone. Best at: rules, alerts and bulk edits that save an hour on Tuesday. Ruled out by: the dashboard is the product, so nobody answers for your margin.

Revealbot runs roughly $79 to $249 a month by spend tier for rules and Slack alerts. Madgicx runs roughly $44 to $249 a month for AI audiences and creative suggestions. Attribution add-ons such as Triple Whale and Northbeam run $99 to $500-plus a month and can still disagree with Ads Manager. Apple’s ATT changes since 2021 made the old tracking model less dependable; another dashboard does not settle the argument.

I used to stack a tool on a small account and call it management. What I bought was a faster way to watch platform ROAS drift from cash. Verdict: cheap help, zero accountability. Ask on the sales call: When platform ROAS and my Shopify payouts disagree, whose number decides whether this worked?

5. Percent-of-spend agencies: more spend, bigger bill

For: ecommerce brands over about $20k a month that want a full team and do not want to manage media day to day. Best at: creative volume and account coverage. A 10-person shop can test more angles than you can alone. Ruled out by: the fee math.

Ten to 20% of monthly spend remains a common agency model; 10% to 20% is also listed for high-growth ecommerce brands spending over $20k a month, usually with a minimum fee. The agency earns more when you spend more, whether that extra spend worked or not. I billed this way early on. A client raising budget from $20k to $30k at 15% paid me another $1,500 while taking on more risk themselves. What the deck calls alignment, I call two people doing opposite jobs.

The reporting can follow the fee. Platform ROAS goes in the slide, blended ROAS goes in the appendix, and margin never makes the meeting. Spend $20k at a 15% fee and you owe $3,000 in management before product and shipping costs enter the picture. Ads Manager can look healthy while the business loses money.

Verdict: real capacity, wrong incentive. Ask on the sales call: Show me a report where platform and blended ROAS disagreed. Which one did you optimize to?

4. Pay-per-result shops: the billable event is the trap

For: a lead-gen business that can define a qualified lead in one sentence and afford to pay for volume while it sorts quality. Best at: putting risk on paper, because you pay when the agreed lead or sale tracks. Ruled out by: the agency has every reason to shape the metric it bills on.

Shops in this lane price risk in at $50 to $200-plus per qualified lead, plus 3% to 10% revenue share on tracked sales. A hybrid base plus bonus is described as the most reliable structure in practice. But the contract cannot rescue a loose definition. Pay per lead and you invite lead volume, not necessarily quality. Pay on last-click ROAS and prospecting becomes easy to starve, leaving last-click ROAS clean while growth stalls.

I have seen this from the inside. The invoice was correct. The pipeline was not. Verdict: useful pricing when you own measurement, dangerous when you do not. Ask on the sales call: I keep ownership of the ad account, pixel and CAPI, and I define a billable lead in writing. Yes or no?

3. Freelancers: straight answers, limited creative hours

For: the small business spending $3k to $10k a month that wants a human who answers the phone and reports what happened. Best at: direct, honest reporting. A good freelancer will tell you blended ROAS fell even when Ads Manager looked fine. Ruled out by: one person’s capacity.

Upwork data puts Meta freelancers at $15 to $40 an hour, with a single campaign setup at $300 to $800 and monthly management at $500 to $2,000. A 2026 rate review gives $25 to $150-plus an hour or $500 to $1,500 a month for one person handling setup and reporting. One person may also be your media buyer, copywriter, editor and analyst. When creative fatigues, testing slows to whatever they can make that week.

I freelanced this lane. Clients got straight answers and slow creative. Verdict: trustworthy but thin. Ask on the sales call: How many new creatives per week, made by whom, and what happens during the two weeks you are on vacation?

2. In-house hire: close to margin, expensive below scale

For: the brand past about $30k a month on Meta where creative testing is a full-time job. Best at: keeping margin decisions close to the business. The person near fulfillment can see returns, stockouts and contribution without waiting for an outside report. Ruled out by: cost below scale.

A specialist with two to three years’ ecommerce experience runs $55k to $75k in salary; a senior hire runs $80k to $100k-plus. With a 20% to 30% burden, a $70k hire becomes $84k to $91k before tools. Add creative, testing tools and training, and the load lands near $95k a year, or about $7,900 a month. That is why I use the rule in who should run your Meta ads at $3k, $10k or $30k a month: test and scale with outside help; bring it in-house when spend can carry the salary.

A salary does not rise every time the ad budget does. But one hire goes on leave, gets poached or never learned creative strategy, and testing stops. I have inherited those accounts: clean pixel, sound structure, nothing tested in six weeks. Verdict: the strongest proximity to margin, with the heaviest fixed cost. Ask in the interview: Walk me through the last time blended ROAS fell while platform ROAS held. What did you kill first?

1. Flat-fee done-for-you: a fixed bill you can test against results

For: the owner spending $3k to $30k a month who wants Meta handled without hiring or paying a larger management bill every time the budget grows. Best at: combining ongoing execution with a fee that does not reward higher spend. Ruled out by: very small budgets, or a business that already has the in-house team to do the work.

groas prices Meta Ads at $1,499 a month per service, regardless of ad spend. It is month to month, with no cut of ad spend, no setup fees and cancellation at any time. Its Facebook and Instagram scope covers audiences, budgets and ads worked around the clock, with new ads tested and winners scaled. A fixed fee still counts toward break-even. It just does not climb because you raise the budget from $10k to $20k.

Here is why I put it first. groas provides an action log with reasoning in plain language and a named human strategist. You can inspect what changed, then check spend against your cleared revenue and contribution numbers. An action log is not a bank statement; you still need your own sales and cost data to judge blended and margin ROAS. That is a better accountability test than accepting a platform screenshot as the verdict.

A percent-fee shop gets a larger bill when spend rises. groas does not. If you want the fee debate in numbers, read percent of spend vs revenue share vs flat fee before your next sales call. Verdict: the best fit here for ongoing Meta management without a spend-linked fee. Ask on the sales call: Can we review ad spend, fee, blended ROAS and margin ROAS together each week, using my revenue and cost numbers?

Who should ignore my #1

Skip the flat-fee pick in three cases:

  • Past about $30k a month, with a full-time creative team already in place: keep the work in-house if you can carry the roughly $7,900-a-month loaded cost and want margin decisions close to fulfillment.
  • Selling high-ticket services with a tight lead definition: a hybrid pay-per-result deal can work if you keep the ad account and tracking and define a billable lead in writing.
  • Under about $3k a month: stay in free Ads Manager and learn to read the vanity metrics your agency uses to distract you before paying anyone.

I left the big brand creative agency out on purpose. Beautiful ads and an awards wall do not make it the right hire when the job is to answer for ROAS. Hire one for thumb-stopping creative alone if that is what you need. I also left out enterprise suites with no public price: if I cannot see the fee, I cannot do the break-even math.

My rule stands: judge a manager by the number they will let you use to judge them. For the owner in the $3k-to-$30k middle, I pick flat-fee done-for-you. If you skipped here, start there. Ask for spend, fee, blended ROAS and margin on one page, every week.