How Facebook ad costs work: CPM, CPC and ROAS unpacked

Facebook has no fixed rate card. You pay an auction clearing price, competing for the same person against every other advertiser who also wants to reach them. So "what does a click cost" has no standard answer.

The useful question is: which factors multiply to set my cost, and which one is easiest to move.

1. The cost formula

Cost per result =  CPM ÷ 1000
                 ÷ CTR
                 ÷ conversion rate

Which is to say:

Metric	Meaning	Mainly driven by
CPM	Cost per thousand impressions	Audience competition, creative quality, account quality
CTR	Clicks ÷ impressions	Creative and copy (the biggest factor)
Conversion rate	Completions ÷ clicks	Landing page, product, price

They multiply. Taking CTR from 1% to 2% halves cost outright, which is usually far more effective than pushing CPM down 10%, and considerably easier to achieve.

Run the numbers once

Say your data looks like this:

	Case A	Case B (creative changed only)
CPM	$30	$30
CTR	1.0%	2.0%
Cost per click	$3.00	$1.50
Conversion rate	3%	3%
Cost per conversion	$100	$50

The only difference between the columns is creative, and cost differs by half. That is why "change the creative first" is not casual advice.

Now compare pushing CPM instead:

	Case A	Case C (CPM down 10%)
CPM	$30	$27
CTR	1.0%	1.0%
Cost per conversion	$100	$90

A 10% improvement, and CPM is mostly not yours to control. It is set by the auction environment, and the part you can influence (creative quality, account quality) improves via creative anyway.

2. The four numbers to watch in practice

In the Ads Manager table, four columns earn their place:

1. Cost per result: the single most important number
2. CPM: whether the auction environment moved
3. CTR (link click-through rate): whether the creative lands
4. Amount spent: whether it is pacing to budget

Everything else can wait a month.

How to read each one

Number	How to read it	What wrong looks like
Cost per result	Against your break-even	Three consecutive days well above break-even
CPM	Against your own last week	Doubling in a day
CTR	The absolute value is informative	Below 0.3% usually means the creative missed
Amount spent	Against budget	Only 20% of budget spent

Only CTR has a broadly applicable absolute benchmark. The other three are comparisons against yourself. This trips a lot of people up: they compare somebody else's CPM against theirs and draw the wrong conclusion.

How many days before you can judge

What you want to judge	Minimum days
Whether money goes out	1
Whether the creative lands	3
Whether cost is acceptable	3–5
Whether to raise budget	5–7
Whether creative has fatigued	7–14

One day of data settles nothing, least of all during learning. The classic beginner error is seeing high cost on day one and changing something, which restarts learning, which raises cost, which prompts another change. The loop never exits.

3. ROAS: whether this business works

ROAS = revenue from ads ÷ ad spend

ROAS 3 means every dollar spent returned three in revenue. That is not the same as profit. Margin decides.

Your gross margin	Break-even ROAS
20%	5.0
30%	3.3
40%	2.5
50%	2.0
60%	1.7
70%	1.4

Work out your own break-even ROAS before reading any ad data. Adjusting ads without that number means not knowing where the target is.

Calculating your break-even ROAS

Break-even ROAS = 1 ÷ gross margin

Gross margin is what proportion remains after direct costs come out of the sale price. Direct costs include:

• Cost of goods or materials
• Payment processing (cards are typically 2% to 3%)
• Shipping and packaging
• Returns

A lot of people deduct only cost of goods and forget payment and shipping. Take a $30 product with $15 of goods cost. That looks like 50% margin. Add 3% payment processing, $2.40 shipping and 5% returns, and the real margin is about 37%, which moves break-even ROAS from 2.0 to 2.7.

That gap is enough to make you scale a campaign that is actually losing money.

One more number beyond ROAS

Lifetime value. If your product gets repurchased, the first transaction's ROAS can sit below break-even, because what you bought was a repeat customer.

The test: work out how many times an average customer buys. If it averages 2.5, you can accept about 40% of break-even ROAS on the first transaction and recover the rest through repeats.

Products without repeats cannot use this. A one-off purchase must break even the first time, or scale simply multiplies the loss.

3b. The auction: who your money is actually competing with

Once this section clicks, every "why" above becomes intuitive.

Advertising is a real-time auction

Every time somebody opens their feed, the system runs an auction to decide whose ad they see. Everyone who put that person in their audience is entered.

But the highest bid does not win. Meta uses a combined score:

Total value ≈ your bid × estimated action rate × ad quality

Three factors:

Factor	Meaning	Can you control it
Bid	What you will pay for this result	✅ Directly
Estimated action rate	The system's odds this person completes your action	🟡 Indirectly, via accumulated signal
Ad quality	User feedback, hide rate, report rate	✅ Through creative

This explains three things

One: why good creative is cheaper. Good creative gives good user feedback and a low hide rate, which raises the ad quality score, which lets the same bid beat a higher one. You are not winning by spending more. You are winning because the system considers your ad worth showing.

Two: why new accounts cost more. A new account has no conversion signal, so the system cannot estimate action rate accurately and defaults to conservative. A conservative estimate means you must bid higher to win, so cost is high. Once signal accumulates the estimate sharpens and the same money buys more.

Three: why a narrow audience is more expensive. Small audience, fewer auctions entered per day, slower sample accumulation, so the system never finds a pattern, so estimated action rate stays inaccurate, so cost never falls. And if other advertisers are chasing that same narrow slice, competition is more concentrated.

Bid strategy: which one for a first run

Strategy	Meaning	Beginner
Highest volume	Spend the budget, get as many results as possible	✅ Use this first
Cost cap	Keep average cost under a value	🟡 You need a baseline first
Bid cap	Never bid above a value	❌ Easy to win nothing at all
Highest value	Maximise total value	❌ Needs value signal

Use highest volume for the first run, because you do not yet know your baseline. Set a cost cap below the going market rate and the ad simply will not deliver, and you will think something else is broken.

Once you have run two full weeks and know roughly where your cost per result sits, consider a cap.

4. When cost rises, which lever to pull

Ranked by effect over cost:

Order	What to change	Why
1	Change the creative	Biggest effect on CTR and CPM, smallest cost
2	Rewrite the first two lines of copy	Phones truncate; those two lines decide whether they read on
3	Check the landing page	Conversion rate is a multiplier; a slow or awkward page destroys everything
4	Adjust the audience	Real effect, but resets learning
5	Adjust budget	Only affects volume, not efficiency directly

Do not change five things at once. You will not know which one worked.

Diagnose the layer before touching anything

Three steps that eliminate most blind fiddling:

1. Did CPM move? Yes → auction environment or account quality, not your creative. Observe three days.
2. CPM flat but CTR dropped? → Creative fatigue. Change creative.
3. CPM flat, CTR flat, but cost per result high? → The problem is the landing page or the product, not the ad.

The most common error is swapping creative when CPM went up. It achieves nothing, because the problem was never there.

What to do in each of the three cases

Case one: CPM up, CTR unchanged. The auction environment changed (a seasonal peak, competitors entering) or account quality declined. Observe for three days while checking Account Quality for violation records. Seasonal means ride it out; an account quality problem means the account is what needs handling, not the ad.

Case two: CPM unchanged, CTR falling. Creative fatigue. The same people seeing the same thing repeatedly respond less. Check frequency (impressions ÷ reach). Frequency clearly rising while CTR falls confirms it. Change the creative, or change the audience.

Case three: CPM and CTR both normal, but conversion cost high. The problem is not the ad. People arrived and did not complete the action. Go look at landing page data: bounce rate, time on page, add-to-cart rate. The ad already did its job.

5. The costs that are not ad spend

Counting only ad spend understates the real number. There is also:

Item	Detail
Asset cost	Pages, BMs, ad accounts. Considerable if your account mortality is high
Downtime	Revenue is zero while an account is disabled. This is the real reason experienced buyers pay more for accounts
Your time	Two hours on an account that was never coming back costs more than the account
Management fees	If outsourced, usually a percentage of spend or a retainer

Run the asset cost with our real prices

Our catalogue has 145 display groups. The price distribution:

	Price (USDT)
Minimum	2
25th percentile	8
Median	10
75th percentile	26
Maximum	200

Mean price is 23.9 USDT.

Say your daily budget is 30 USDT and you replace two accounts a month at the median price:

Monthly ad spend  = 30 × 30 = 900 USDT
Monthly asset cost = 10 × 2  =  20 USDT
Asset cost share   = 2.2%

Only 2.2%, which looks unimportant. But at two accounts a week:

Monthly asset cost = 10 × 8 = 80 USDT
Asset cost share   = 8.9%

And with downtime, the real impact is well beyond 8.9%. Every replacement means warming up again and running the learning phase again, and that stretch is your least efficient spend.

Downtime is the real number

Daily budget	Revenue lost per day down (at ROAS 2)	Price gap to a better account
$10	$20	$3–10
$30	$60	$3–10
$100	$200	$3–10
$300	$600	$3–10

The upgrade test is simple: when a day of downtime costs more than the price gap to a better account, upgrade.

Somebody spending $10 a day buying the most expensive account is wasting money; somebody spending $200 a day saving 10 USDT has their priorities inverted. Differences between tiers are in account types and when to use them.

Bundles versus buying separately

Ignore "save 30%" claims. Compare parts total against bundle price:

Bundle	Bundle price	Parts total	Saved	Discount
First run	39 USDT	48 USDT	9 USDT	18.8%
Account died again	49 USDT	62 USDT	13 USDT	21.0%
Local market	89 USDT	108 USDT	19 USDT	17.6%
Steady running	129 USDT	143 USDT	14 USDT	9.8%

The steady-running bundle has the smallest discount (9.8%) but contains the most expensive items. A low discount percentage does not mean poor value. What matters is which pieces you were going to buy anyway. If you only need two of them, separate is cheaper.

6. Why new accounts start expensive

Three factors compound:

1. Low spending limit: you cannot get volume, so no scale efficiency
2. No conversion signal: the system does not know who to show it to, so it guesses
3. The learning phase: every significant change restarts it

So the right move on a new account is not spending hard until it gets cheap. It is producing a record first: small budget, few changes, letting signal accumulate. Full rhythm in the gradual scaling curve.

What the first four weeks look like

Period	Relative cost	What you do
Days 1–3	Highest	Change nothing, just confirm money goes out
Days 4–7	Starting to fall	First judgement; decide about creative
Days 8–14	Clearly improving	Start raising budget along the curve
Days 15–28	Settling	Now the numbers are your baseline

Do not treat week one's cost as a verdict on the product. That number is necessarily ugly, because the system is spending money to buy information.

6b. Following along: the full arithmetic on a $30 product

Formulas above, a worked example here. This is the fastest way to see where money leaks.

Step 1: work out the real margin

Say you sell a $30 product:

Item	Amount	Note
Price	30.00
Cost of goods	−12.00	Purchase or production
Payment processing	−0.90	About 3%
Shipping and packaging	−2.40
Returns	−1.50	About 5%
Gross profit	13.20
Margin	44.0%

Deduct only cost of goods and you would calculate 60% margin. That gap magnifies in the next step.

Step 2: break-even ROAS

Break-even ROAS = 1 ÷ 0.44 = 2.27

Using the wrong 60% figure gives 1.67. You would scale a campaign at ROAS 2.0 believing it profitable, while every order loses money.

Step 3: convert to a cost ceiling per conversion

Max ad cost per conversion = 30.00 ÷ 2.27 = $13.20

So as long as acquiring one order costs less than $13.20 in ads, you are ahead.

Step 4: work back to a cost ceiling per click

Say your landing page converts at 3%:

Max cost per click = 13.20 × 3% = $0.40

That is the number to watch daily. It is far more concrete than "what ROAS do I need", because Ads Manager has that column already.

Step 5: check feasibility

Now look at your actual data:

Your actual CPC	Verdict	What to do
Below $0.40	Profitable	Raise budget along the curve
$0.40 to $0.60	Near break-even	Improve CTR or the landing page first
Above $0.60	Losing money	Stop and rethink; do not raise budget

Step 6: find the most effective improvement

Same target (cost per conversion from $18 down to $13.20), three routes:

Improve which	By how much	Difficulty
CTR from 1% to 1.4%	+40%	Moderate; a new creative can do it
Conversion rate from 3% to 4.2%	+40%	High; means changing the landing page and product
CPM from $30 to $21.50	−28%	Very high; mostly outside your control

The three routes have identical effect and wildly different costs. That is the mathematical basis for "change creative first", not a rule of thumb.

Three common errors in this calculation

One: deducting only cost of goods. Payment, shipping and returns all come out. The gap was shown above.

Two: using average order value instead of this product's price. With several products, calculate separately, because margins differ.

Three: forgetting repeats. If customers average 2.5 purchases, the first transaction can carry a higher acquisition cost. But this needs real data, not a hoped-for number.

7. How to read ad management fees

If you are considering outsourcing, this section is for you.

The three common fee structures

Structure	How it works	Favours you when
Percentage of ad spend	A fixed percentage of what you spend	Spend is small
Fixed retainer	A set amount monthly	Spend is large
Retainer plus performance	A base plus a results bonus	Both sides want to tie it to outcomes

The thing to watch with percentages: the fee grows proportionally as your spend grows, while the workload does not necessarily grow with it. Running $3,000 a month and $30,000 a month are not ten times apart in operational complexity.

Three things to know yourself before outsourcing

1. Your break-even ROAS. Without it, you cannot judge whether they are doing well
2. Your current baseline. Without a baseline, any number can be described as "already good"
3. Who owns the account. The most important one, expanded below

Which permission to grant

Advertiser, not Admin.

Admins can remove other admins, including you. It has happened that an engagement ended and the other party never removed themselves, or removed the original owner.

Advertiser permission covers the full job but cannot move your assets. See the permissions section in what is an ad account.

One more thing: the assets should sit under your BM, not the agency's. When the engagement ends, you want to remove a person and keep running, not negotiate to get your assets back.

7b. Building your own baseline

The piece repeatedly says "compare against yourself". This is how you build that self.

Why a baseline is mandatory

Without one you judge by somebody else's numbers or by feel, and both mislead:

• Their CPM reflects their vertical, region and season, not yours
• Feel makes you overreact to normal variance and underreact when something is genuinely wrong

With a baseline, "is this number normal" becomes a lookup rather than a judgement call.

How to build it

Run four full weeks, recording three numbers a day. At the end you have:

Metric	What you record	Used for
CPM	Average and range	Judging whether the auction environment moved
CTR	Average	Judging creative
Cost per result	Average, and your best three days	Judging whether to raise budget
Frequency	Which day it starts clearly rising	Judging how fast the audience wears out

One spreadsheet is enough. The columns are date, spend, impressions, clicks and results; everything else is a formula.

What the sheet looks like

Date	Spend	Impr.	Clicks	Results	CPM	CTR	Cost/result
08-01	5.00	5,000	55	2	1.00	1.1%	2.50
08-02	5.00	4,800	60	3	1.04	1.25%	1.67
08-03	5.00	5,200	48	2	0.96	0.92%	2.50

Record the first five columns; the last three are formulas. This sheet's value in three months is far above the ten minutes it costs now.

How often to refresh the baseline

Quarterly. The auction environment moves with the seasons, and judging November's data against March's baseline distorts things, especially around shopping peaks.

Three other events force a rebuild:

• The product or price point changed
• The main audience changed
• The account changed (a new account's cost curve differs from a mature one's)

7c. Four cost quirks outside the US market

Things English tutorials skip and non-US advertisers always hit.

① The hidden cost of currency

On a USD account your bank typically adds about 1.5% in foreign transaction fees. At $1,000 a month in spend, that is $180 a year.

It never appears in Ads Manager, only on your card statement, which is why most people never count it.

② Seasonal auction pressure

Retail calendars are highly concentrated in most markets. During the big shopping events, the same audience is chased by a lot of advertisers at once and CPM rises noticeably.

Two responses:

• Get budget and creative ready two weeks before the peak, rather than starting to test during it
• Rising cost during a peak is normal; do not judge it against your ordinary baseline and conclude the creative broke

③ The ceiling on a small-language audience

Smaller-language markets have a hard ceiling on reachable people. Two consequences:

1. Audiences wear out fast, and frequency rises faster than in large markets
2. "Just widen the audience" runs out of road quickly

So in a small market you change creative more often than you would in the US. The same people see you repeatedly, and fatigue arrives sooner.

④ Invoices and tax

Meta ad receipts download from the Billing page. Two things people always ask:

• The issuing entity is usually one of Meta's overseas entities, not a local company
• What downloads is a receipt, which is not the same as a locally compliant tax invoice in many jurisdictions

This runs into tax classification and accountants handle it differently. It is outside what we can advise on; ask yours.

8. A common misdiagnosis

"My CPM is higher than my competitors', so something is wrong with my account."

Usually not. CPM swings widely with region, vertical, season and that week's competition, so cross-account comparison is close to meaningless. The meaningful comparison is your own CPM last week.

If your own CPM doubles within a few days, that is worth investigating. Usually creative fatigue, an exhausted audience, or a shift in the overall auction environment such as a seasonal peak.

Why cross-account comparison fails

Variable	Effect on CPM
Vertical	Large. Finance, insurance and retail are completely different auctions
Region	Large. The same product differs several-fold between countries
Season	Large. During shopping peaks everyone is chasing the same placements
Audience size	Medium. Very narrow audiences usually carry higher CPM
Creative quality	Medium. The system rewards better-performing creative with lower cost
Account quality	Medium. Accounts with violation records get marked up

The first three you cannot control, and they matter most. So comparing CPM with somebody else compares your different verticals and timing, not who is better at this.

9. Five myths that cost money

"A higher budget makes the system try harder." Budget decides what you are willing to spend, not the system's effort. Too high a budget on a new account trips risk controls instead.

"Lower CPM is always better." Not necessarily. A very narrow audience sometimes has low CPM, and those people never buy. Watch cost per result, not one link in the chain.

"ROAS 3 means profit." Margin decides. At 20% margin, ROAS 3 loses money; break-even is 5.0.

"High cost means a bad account, so buy a pricier one." Nine times out of ten high cost comes from creative and the landing page, not the account. A more expensive account does not improve CTR.

"Running several ads at once finds the winner faster." Too many at once compete for the same audience, budget gets divided, and none of them finishes learning. One or two is enough for the first month.

9b. How the learning phase affects cost

The piece keeps mentioning learning. This section finishes the relationship.

Cost during learning is necessarily higher

The system starts out not knowing who inside your audience responds. It shows the ad to a small group, watches who clicks and who converts, and adjusts. That probing costs money, and that money lands in your cost.

So high cost during learning is not an anomaly. It is inevitable. The only question is whether learning ever ends.

What restarts it

Action	Restarts learning	Cost impact
Large budget change (past 50%)	✅	High
Audience change	✅	High
Placement change	✅	High
Optimisation event change	✅	High
Pausing and resuming	✅	High
Creative swap	🟡 Smaller effect	Medium
Copy edit	🟡 Smaller effect	Medium
Renaming the ad	❌	None

Pausing and resuming restarts learning, and almost nobody knows this. Plenty of people switch ads off at the weekend to save money, switch them back on Monday, then complain that cost went up. What they are doing is resetting the learning phase every weekend and paying the probing cost over and over.

The arithmetic: switching off two days a week means four learning restarts a month. If each restart raises cost by 30% for three days, that is 12 days a month at elevated cost, which is 40% of the time.

The budget saved by pausing at weekends is far below the cost of repeatedly restarting learning.

Three ways learning never finishes

1. Audience too small: daily impressions are limited, so samples accumulate too slowly
2. Budget too low: same mechanism
3. Changing too often: every change restarts it, so it never completes

The third is the most common among beginners, and it is self-inflicted.

Which is why "change nothing for three days"

See high cost on day one, change something, learning restarts, cost rises, change again. You stay inside the learning phase permanently, paying the highest possible cost.

The right move: record for three days and touch nothing. After three days you have a curve, and a curve can be judged.

9c. Turning cost control into a weekly routine

Understanding the mechanism is not enough. You need a rhythm. Copy this one.

Daily (about 3 minutes)

1. Set the date range to yesterday
2. Record three numbers: spend, cost per result, CTR
3. Check whether any ad flipped to rejected or error

What you do not do: change settings because one day's numbers looked bad.

Weekly (about 20 minutes)

1. Set the date range to the last 7 days
2. Add CPM and frequency, compared against last week
3. Run the three-step diagnosis (CPM → CTR → landing page)
4. Decide about raising budget (never above 50%, at least 48 hours apart)
5. A sustained CTR decline means preparing the next creative set

Monthly (about 40 minutes)

1. Reconcile: Ads Manager's monthly spend against the bill
2. Recalculate your real margin, because costs move
3. Update your break-even ROAS
4. Check asset cost share (how many accounts you replaced this month)

Item 2 is the one people skip. Purchase costs, shipping and payment rates all move, and your break-even ROAS is built on margin. Margin changed and you did not recalculate means judging new data by an old standard.

The point of this table is restraint, not diligence

Beginners' problem is usually not looking too little. It is looking too often and wanting to act every time. Three minutes a day recording three numbers beats refreshing the report hourly, because the first accumulates a curve you can judge and the second only makes you decide inside noise.

9d. Three business types, three cost structures

"How do ad costs work" has a different answer depending on what you sell. These three are the most common.

E-commerce: watch ROAS

The result you want: orders.

The key number: break-even ROAS, calculated above.

What is distinctive: there is a specific revenue figure per order, so ROAS applies directly. The things to be careful about are calculating margin correctly and whether repeats exist.

Where it usually goes wrong: deducting only cost of goods, forgetting payment, shipping and returns. As shown, that gap is enough to make a losing campaign look profitable.

Lead generation (insurance, property, courses, B2B): watch cost per lead

The result you want: contact details.

The key number:

Acceptable cost per lead = average deal value × margin × lead close rate

Example: $1,000 average deal, 40% margin, 5% close rate:

1,000 × 40% × 5% = $20

So a lead costing under $20 in ads is profitable.

What is distinctive: ads only get you to "lead acquired". After that come follow-up, quoting and closing. So the close rate is your capability, not the ad's responsibility.

Where it usually goes wrong: using a hoped-for close rate rather than a measured one. Somebody doing this for the first time does not know their close rate, so they have to run a batch of leads to measure it. That batch is tuition.

App promotion: watch post-install behaviour

The result you want: installs, but what you actually care about is whether they use it and pay.

The key numbers: cost per install, plus retention and lifetime value.

What is distinctive: installs are cheap and churn fast. Watching only cost per install buys you a pile of people who installed and deleted.

Where it usually goes wrong: optimising for "install" instead of a key post-install action. The system is obedient. Ask for installs and it delivers installs, regardless of whether those people ever open the app again.

What the three share

	E-commerce	Lead gen	App
Where the ad's job ends	The order	The lead	The install
Where your job starts	Landing page and product	Follow-up and closing	Retention and monetisation
Most common misjudgement	Margin calculated too high	Close rate is a hope	Only watching install cost

The common thread: an ad's responsibility has a clear boundary, and problems beyond that boundary cannot be fixed by adjusting the ad. This is the root cause of people adjusting endlessly with nothing to show.

10. Quick glossary

Term	In plain terms	Formula
CPM	Cost per thousand impressions	Spend ÷ impressions × 1000
CPC	Cost per click	Spend ÷ clicks
CTR	Click-through rate	Clicks ÷ impressions
Conversion rate	Share of clickers who complete the goal	Conversions ÷ clicks
Cost per result	What one of your specified results costs	Spend ÷ results
ROAS	Return on ad spend	Revenue ÷ ad spend
Break-even ROAS	The ROAS needed to break even	1 ÷ gross margin
Frequency	How many times one person saw it	Impressions ÷ reach
LTV	Lifetime value	Average transaction × average repeat count

10b. When to stop

Possibly the most money-saving section here, because a beginner's most expensive loss is rarely getting it wrong. It is getting it wrong and continuing to raise budget.

Three signals to stop

One: five consecutive days with cost per result above 1.5× break-even.

Not three days, because three is still inside learning-phase variance. Five days consistently high is structural, not noise.

Once stopped, run the three-step diagnosis (CPM → CTR → landing page) and find the layer before touching anything.

Two: CTR under 0.3% after two creative sets with no improvement.

That usually means your message does not match your audience, and new creative will not save it. What needs rethinking is who you are talking to, not which image to use.

Three: conversion rate under 1% while the traffic itself is right.

People arrive, time on page is normal, and they do not buy. The problem is the product, the price or the page. Continuing to advertise just delivers more people to somewhere that cannot convert.

What to do after stopping

Do not switch it off and start over. Follow this order:

1. Drop budget to minimum, do not stop entirely (stopping breaks your spend record and resets learning)
2. Extract the data: best creative, best audience, best time of day
3. Change one variable, usually creative
4. Hold the low budget for three days and look for improvement
5. Improvement means climbing back along the curve; no improvement means the next variable

The cost everyone forgets

Your own time.

Say your time is worth $15 an hour. An ad set that will not work, with an hour a day spent watching it, changing it and thinking about it, is $450 a month in time cost that appears on no report.

Set a time budget: "five more hours on this, and if it still will not run I change the subject." Writing your stop-loss as a number beats deciding by feel.

Stopping is not failure

An ad set that never worked tells you four things, provided you recorded them:

• This audience does not respond to this message
• Where this creative's CTR lands
• Roughly what your landing page converts at
• What range your actual CPM sits in

Those numbers are your baseline. The next campaign starts from data rather than guesswork, and that is what the money bought.

11. Three numbers to calculate before you start

• My gross margin (with payment, shipping and returns deducted, not just cost of goods)
• My break-even ROAS (1 ÷ margin)
• Whether my customers repeat, and how often (if so, the first transaction can carry a lower ROAS)

Until those three exist, reading an ad report is pointless, because you do not know what "good" means.

---

In one sentence: calculate break-even ROAS, watch cost per result, and change creative first when cost rises. Everything else is secondary.

---

Sources (all official pages, checked 2026-09-08). Prices, catalogue distributions and warranty ratios in this article are our own catalogue and ticket data, not taken from these pages:

• How much it costs to advertise on Meta — Meta Business Help Centre — facebook.com
• How the ad auction works — Meta Business Help Centre — facebook.com
• About the learning phase — Meta Business Help Centre — facebook.com
• Meta ads budget and pricing — Meta for Business — facebook.com
• Advertising Standards — Meta Transparency Center — transparency.meta.com
• Ad metric field definitions (Marketing API Insights) — Meta for Developers — developers.facebook.com