Facebook ads can be worth testing when the customers you acquire contribute more than the cost of reaching, converting and serving them. They can also lose money while producing attractive clicks, cheap leads or a positive reported ROAS. Start with your offer and economics, then decide what evidence an affordable pilot can realistically provide.
A Quick Decision: Is Facebook Worth Testing for You?
| Situation | A sensible next step |
|---|---|
| Clear offer, known margins and working checkout | Plan a limited purchase-focused pilot |
| Local service with a defined service area | Measure reachable, qualified enquiries and actual customers |
| Long sales cycle or B2B offer | Connect lead stages and follow-up before judging raw CPL |
| Low margin and uncertain repeat purchase | Resolve the allowable acquisition cost before spending |
| Broken page, weak offer or no outcome tracking | Fix the missing foundation first |
| Very limited test budget | Reduce the question and scope; allow an inconclusive outcome |
These are decision prompts, not predictions that a business category will succeed. A high-margin product can still have poor demand, and a low-priced offer can work when measured contribution and acquisition cost support it.
Calculate What a Customer Can Actually Contribute
Start with net revenue after relevant discounts and refunds, then subtract the variable costs required to deliver the sale. Include product cost, fulfilment, payment fees, expected returns and other applicable variable costs once. Decide how much must remain for overhead, risk and profit.
Illustrative first-order example: a shop receives $100 net revenue from an order and incurs $60 of non-ad variable costs. It has $40 before acquisition cost. If it wants $10 left after advertising, the target acquisition cost is $30.
| Illustrative first-order outcome | Amount |
|---|---|
| Net order revenue | $100 |
| Non-ad variable costs | $60 |
| Contribution before advertising | $40 |
| Required remainder | $10 |
| Target acquisition cost | $30 |
| ROAS at that target | 3.33x |
At $40 acquisition cost, this example reaches first-order contribution break-even before overhead and other costs. At a $30 target, it leaves $10. The arithmetic is not a benchmark or an estimate of what Facebook will charge.
If repeat purchases matter, use measured contribution within an explicit payback window and account for cash flow. Do not multiply revenue by an optimistic lifetime-value estimate and treat it as money already available to fund acquisition.
Why Positive ROAS Does Not Necessarily Mean Profit
ROAS divides attributed revenue by ad spend. It does not subtract product cost, returns, fees, labour or overhead. An aggregate industry ROAS cannot establish that the median advertiser is profitable or that your business can afford the same CPA.
When the contribution-margin definition already includes all relevant non-ad variable costs, contribution break-even ROAS is 1 divided by that margin fraction. At 40%, it is 2.5x before fixed costs and the required profit reserve. Keep revenue and cost definitions consistent.
Use our ROAS and break-even calculator or customer acquisition cost calculator as arithmetic aids. A calculator does not forecast demand or prove incremental sales.
For Local Services: A Cheap Lead Is Not the End Result
Suppose a fictional service business spends $800 and receives 80 enquiries. Twenty are reachable and qualified; four become customers. Raw CPL is $10, cost per qualified lead is $40 and ad spend per acquired customer is $200. Sales and fulfilment costs still need accounting.
If a completed job contributes $150 before acquisition, that campaign does not work on those assumptions. If it contributes $400 and the follow-up cost is affordable, it may be worth further investigation. These are hypothetical scenarios, not typical local-service results.
- Define “qualified” before looking at performance.
- Record contactability, service location, customer need and agreed sales stage.
- Review how quickly and consistently enquiries are followed up.
- Compare outcomes after the relevant sales delay, not only on the day the form arrives.
Spam, wrong numbers and mismatched demographics are reasons to inspect the actual ad path and settings. They are not proof of one universal platform fault. See our wrong-audience diagnostic.
When Facebook May Be a Useful Channel
It can be a candidate when you can present a relevant offer to people who were not actively searching for it, demonstrate a product or explain a problem visually. A clear destination and a credible next step matter as much as an attention-grabbing asset.
Search advertising can be a candidate for existing expressed demand. Email, referrals and organic activity can serve other roles. Compare qualified outcomes and complete costs rather than declaring one channel universally cheaper from unrelated benchmark samples.
A boosted post and an Ads Manager campaign should also be assessed by objective and outcome. Engagement is useful when engagement is the intended goal; it is not automatically evidence of profitable customer acquisition.
What Meta’s Automation Does and Does Not Settle
Meta’s current Advantage+ sales guidance describes automation across audience, placements and budget. It recommends first-party data and diverse creative, while explicitly noting that a high opportunity score does not guarantee performance.
Those tools do not remove the need to define a valuable customer, verify measurement and approve the offer. A creative refresh every two weeks, broad targeting for every business or a particular number of live ads is not a universal formula for success.
Design an Affordable Pilot Around One Question
Write the question first: can this offer attract qualified purchases or enquiries at a cost the business can support? Choose a budget you can afford to spend on learning and an explicit point at which you will review the evidence.
| Pilot decision | Write down |
|---|---|
| Offer | Product or service, price and intended customer |
| Outcome | Purchase or clearly defined qualified lead |
| Economics | Target acquisition cost and maximum affordable test loss |
| Setup | Working destination, appropriate controls, identity and measurement |
| Creative | A manageable set of distinct hypotheses with truthful evidence |
| Timing | Review window that includes relevant conversion or sales lag |
| Decision rule | Continue, revise, stop or remain inconclusive |
There is no universal minimum budget or mandatory three-month commitment. A small pilot may answer only operational or directional questions. If the available information cannot distinguish a useful outcome from noise, say so rather than treating no clear answer as proof the channel cannot work.
Measure Beyond the Platform Dashboard
Compare ad-platform reporting, analytics and the business record using aligned dates, currencies, time zones, attribution windows and event definitions. They can disagree without one being broken. Check consent, redirects and duplicate or missing events where relevant.
Google Analytics attribution settings assign credit under configured models and windows. Assigned credit is different from proving that an ad caused an additional sale. A suitable lift or holdout experiment is a separate measurement question.
Keep the pilot’s operational evidence as well: which ads were approved, which destinations worked, how many enquiries were reachable and what follow-up actually happened. These can explain poor outcomes more usefully than a single blended ROAS.
When Software Becomes Relevant
If the business economics and offer work but campaign preparation is a recurring bottleneck, evaluate software against that task. AdManage supports repeatable launch operations with approved media and structured naming and tracking workflows.
Measure preparation, review and exception-handling work against the current plan cost. An operating-efficiency improvement is different from proof that the software generated more profitable sales.
Are Facebook Ads Worth It? FAQs
Are they worth it for a small business?
Potentially, when the offer, service capacity, contribution and qualified acquisition cost fit. Use an affordable pilot instead of relying on a category-wide yes or no.
How much should I spend to find out?
Choose the question, available information and maximum affordable loss first. A fixed daily amount does not guarantee a useful conclusion, especially with a low conversion rate or long sales cycle.
Is 2x ROAS good?
It depends on margins, costs, payback and attribution. At a 40% contribution margin before advertising, 2x is below contribution break-even on the stated revenue basis.
Why are my leads cheap but unprofitable?
Inspect qualification, contactability, sales conversion and follow-up costs. Raw CPL measures only one stage of the process.
Should I keep testing forever?
No. Use a pre-agreed affordable budget, review window and decision rule. Continue when the evidence and business case support it; revise or stop when they do not.
About the author
Cedric Yarish
Co-founder of AdManage.ai
