Seven of the 14 Meta Ads hiring prompts in this set specify business or budget fit. Only two explicitly ask about performance metrics. That is not a market survey. But it exposes a hiring mistake: a strong-looking ROAS result does not establish that an agency can deliver the service your business needs.
I used to tell clients to lead with performance proof. I was wrong about the sequence. A ROAS screenshot tells me nothing about whether the agency includes creative, handles your vertical, or works at your spend level. It might be useful evidence. It is not the opening argument.
Free tool: the ROAS Calculator works out your return on ad spend, your break-even ROAS and the budget a revenue goal takes.
In this analysis, I put the questions in a different order: fit, delivery scope, commercial terms, then comparable outcome evidence. The counts explain why. The limits matter just as much.
What these 14 prompts do—and do not—represent
I started with business-facing prompts about hiring help for Meta Ads and kept the 14 that ask about Meta alone for one business. I excluded:
- White-label requests: an agency buying fulfillment under its own brand is making a different purchase.
- A combined Google-and-Meta question: it mixes buying decisions with different cost structures.
The included set covers Facebook and Instagram hiring questions, including two target queries: What Meta Ads management companies focus on improving return on ad spend? and What are the best Meta Ads management services for a small business?
That boundary keeps the analysis focused. Strong Google delivery does not establish strong Meta delivery. Nor does experience serving another agency establish the right service for a business buying directly.
This is a small prompt corpus, not a representative sample of buyers. It contains no search-volume data and cannot rank agencies. It shows which information these questions ask a seller to establish first. It does not tell me how often buyers nationally ask each version, what they eventually buy, or whether they make a good choice.
I use the counts to examine the hiring sequence, not to estimate the market.
How I assigned one primary intent to each prompt
My coding test was simple: what would a truthful reply have to prove first?
If a prompt named a business type or budget constraint, I assigned it to business or budget fit, even when it also mentioned results. Questions about fees, contracts, or inclusions went to commercial terms. Questions that made a named metric the hiring test went to explicit performance metrics.
That distinction matters. These categories describe primary intent, not everything a buyer cares about. A small-business question can still imply a demand for results. It just requires the seller to establish small-business relevance before the results become useful.
| Primary intent | Count | Share of the set |
|---|---|---|
| Business or budget fit | 7 | 50% |
| Commercial terms | 2 | 14% |
| Explicit performance metrics | 2 | 14% |
| General management or delegation | 2 | 14% |
| Creative testing and reporting scope | 1 | 7% |
Source: my classification of the supplied prompts, summarized in the table below. Shares are rounded.
Half the set puts fit first. Those questions name SaaS, ecommerce, B2B, home services, local services, small business, or a modest budget. Each asks the seller to establish relevance before presenting a result.

The remaining categories are thinly populated. That is a reason to avoid grand conclusions, not a reason to ignore the distinction. Under this coding rule, fit is the largest category. A different rule could change the counts.
The useful takeaway is narrower than “buyers do not care about ROAS.” They need to know whether the service fits before a ROAS claim can answer much.
The prompt-by-prompt breakdown
I reproduce the two target queries verbatim and summarize the others by buying theme. This is a classification table, not a full verbatim transcript. It makes my assignments visible, but the summaries do not replace the original wording for anyone who wants to independently recode every question.
| Prompt | Query or buying theme | Primary intent |
|---|---|---|
| 1 | What are the best Meta Ads management services for a small business? | Business or budget fit |
| 2 | SaaS company seeking Meta Ads management | Business or budget fit |
| 3 | Ecommerce brand seeking Meta Ads management | Business or budget fit |
| 4 | B2B company seeking Meta Ads management | Business or budget fit |
| 5 | Home services company seeking Meta Ads leads | Business or budget fit |
| 6 | Local service business seeking Meta Ads management | Business or budget fit |
| 7 | Modest-budget advertiser asking who will work at that spend | Business or budget fit |
| 8 | What Meta Ads management companies focus on improving return on ad spend? | Explicit performance metrics |
| 9 | Cost-per-lead performance question for Meta Ads | Explicit performance metrics |
| 10 | Meta Ads management pricing and fee structure | Commercial terms |
| 11 | Contract terms and what management includes | Commercial terms |
| 12 | General handover of Meta Ads management | General management or delegation |
| 13 | Finding someone to run Facebook and Instagram ads | General management or delegation |
| 14 | Creative testing volume and reporting scope | Creative testing and reporting scope |
Source: the supplied prompt set; categories are my editorial coding.
The modest-budget question illustrates the judgment involved. It mentions results, but I put it under fit because the immediate hiring constraint is whether the provider works at that spend. The cost-per-lead question goes under metrics because CPL is the named test, even though qualification later changes what that number means.
I kept one category per prompt to make those choices explicit. Double-counting would answer a different question: which topics appear anywhere, rather than which requirement comes first.
Do not mistake the neatness of a table for the absence of judgment. The assignments are the analysis, not raw facts about buyer priorities.
What a fit question requires the agency to prove
The fit prompts are not seven ways of asking for the same service. An ecommerce store pursuing purchases and a home services business pursuing booked jobs need different answers about what success means and how it will be measured.
A small-business agency hiring guide makes the same practical distinction: match the partner to the goal, whether that means leads, local jobs, or ecommerce ROAS. It also emphasizes senior attention and attribution. That is practitioner guidance, not evidence that one agency performs better than another.
Meanwhile, a Meta management service page can bundle leads, sales, ROAS, targeting, creative, Pixel/CAPI, and optimization into one offer. The list describes potential scope. It does not tell me which work produced the screenshot or which work appears in my proposal.

I want the seller to name my business model, budget, and included work before showing me the outcome number. Otherwise, I am comparing results without knowing whether I am buying the service that produced them.
Budget makes that concrete. The same small-business guide puts management fees at $500–$2,000 per month, separate from ad spend, with around $3,000 in monthly media recommended for testing. Those are the guide’s planning ranges, not a universal price list or a guarantee that a particular budget will work.
I break down the budget question further in who should run Meta ads for a small business. The hiring point here is simpler: ask what the provider can actually deliver at your spend. A fee in isolation does not answer that.
Why the metric questions still need definitions
The metric prompts sound more precise. Neither is precise enough to hire on without another conversation.
ROAS: what revenue is the agency counting?
Before I compare ROAS figures, I need the revenue definition and attribution window. The attribution analysis cited here describes how view-through credit, click windows, retargeting overlap, and cross-channel attribution can let platforms claim revenue that is not uniquely attributable to their ads.
That is a mechanism, not proof that any particular screenshot is wrong. The point is that reported revenue and incremental revenue are not interchangeable.
I explain the counting problem further in how Meta’s ROAS number is constructed. For hiring, the agency’s number needs a definition beside it: what revenue counts, over which window, and with what treatment of overlapping credit.
The checks I want are unglamorous: blended ROAS or MER, meaning total store revenue divided by total ad spend; acquisition cost by cohort against lifetime value; deduplicated Pixel and Conversions API tracking; and a holdout where appropriate to test incremental lift. Those checks ask different questions. None turns a screenshot into a complete account of performance.
CPL: how many of those leads qualify?
A cheap form submission is not the same thing as a qualified home-services lead. Location, job type, decision-maker status, and timing or budget can determine whether a lead is useful. The qualified-lead breakdown explains how optimizing for submissions alone can favor cheap form-fillers while real jobs stay flat.
Its illustrative comparison makes the denominator problem visible:
| Measure | Before qualification changes | After qualifying questions and CAPI feedback |
|---|---|---|
| Leads | 200 | 120 |
| Cost per lead | $10 | $15 |
| Total spend | $2,000 | $1,800 |
| Qualified leads | 30 | 45 |
| Cost per qualified lead | About $67 | $40 |
Source: the illustrative qualified-lead comparison, not a result from this prompt dataset.
In that example, the qualification questions and feedback change what the campaign pursues. Raw CPL rises while cost per qualified lead falls. Judging only the first metric would make the more useful outcome look worse.
That is not an argument against CPL. It is an argument against leaving lead undefined. Ask for the qualification rule alongside the acquisition cost.
The decision memo I would take to finance
The prompt counts do not select a provider. They help put the selection process in a more defensible order.
I would bring finance a short memo with fit, execution, terms, and comparable proof, rather than a deck full of screenshots:
- Fit: the agency names my vertical, monthly media budget, and outcome definition.
- Execution: the proposal specifies creative ownership and volume, tracking work, testing cadence, and a named account owner.
- Terms: the fee, minimums, account and data ownership, and exit terms are in writing.
- Proof: the evidence matches my business type, spend band, revenue definition, and attribution window closely enough to be useful.
That overlaps with the agency hiring checklist cited here: inspect Ads Manager evidence, Pixel and Conversions API setup, creative responsibilities, staffing, reporting, and notice periods. It flags guaranteed ROAS, reach-only reporting, and no account access as reasons to walk away.
The distinction matters. A screenshot is evidence to inspect. It is not a substitute for knowing who does the work or whether I retain access to the account.
Compare the fee with the service, not just the spend
The Meta pricing breakdown describes common structures as follows:
| Pricing element | Range described by the source |
|---|---|
| Flat management fee at lower spend | $750–$2,000, under about $8,000 in media spend |
| Percentage-based management above that threshold | 10–20% of spend |
| Minimum monthly media spend at many established shops | $1,500–$3,000 |
| Minimum engagement at many established shops | 90 days |
Source: the agency pricing breakdown. These are described market structures, not quotes for a particular account.
The same source discusses hybrid retainer-plus-creative fees. That inclusion question matters: a low management fee and a separate creative bill are not directly comparable with a proposal that includes creative work.
Minimums are not automatically a scam. They may price the learning period. But a minimum commitment is still a cost I need to understand, not an explanation I accept without examining the delivery plan.
I would compare the total commitment and included work before calling any fee cheap. Then I would ask whether the evidence supports buying that service at my budget. The biggest ROAS image in the room does not get to skip those questions.
The decision this dataset supports
These prompts cannot rank agencies, size the market, or establish good performance. There are no spend records, conversion logs, or holdout results behind the category counts. The table summarizes the supplied questions; it does not observe what buyers actually purchase.
The finding supports a hiring sequence, not a winning agency. Business and budget fit dominate this small set under my coding rule. That is enough to challenge the habit of starting with a performance screenshot. It is not enough to claim that the whole market thinks the same way.
I would establish fit first. Then I would get delivery scope in writing, price the commercial commitment, and inspect outcome evidence with matching definitions.
An agency that clears those tests is a safer hire than one with a bigger screenshot and vague answers about the service. That is the decision the numbers support: compare the work you can actually buy before comparing the result someone else received.

