What a Meta Ads Agency Does, and How to Choose One

A Meta ads agency runs three jobs: the creative that gets made, the money spent behind it, and the measurement that says whether any of it worked. Hayes Media does all three, with creative and media buying under one roof, and the work is pointed at one outcome: helping eCommerce brands scale past $10M a year. Hire one when your constraint has stopped being account settings and started being creative volume.

What a Meta ads agency actually does

Most brands think they're buying media buying. They're buying a creative supply chain with a buying function bolted to the end.

The work splits into six jobs.

• Creative strategy. Deciding what to say before anyone films anything: the angles, the offers, the objections, and which customer belief each ad is built to move. Skipping it is why accounts produce volume without winners.
• Creative production. Sourcing creators, scripting, shooting, editing, cutting the variations. Our homepage describes it as "sourcing content creators, ideation, scripting, and editing." Format range in Facebook ad creative best practices and UGC ads examples.
• Media buying. Account structure, budgets, bid strategy, exclusions, and the calls about when to push spend and when to stop. More in scaling Facebook ads.
• Testing. The system that decides what goes live, what gets killed, and what gets iterated on. Ours is written up in the creative testing framework.
• Landing and offer. The page the click lands on and the offer it carries. An agency that refuses to touch either is optimizing half a funnel.
• Measurement. The number that says whether the program is winning, and who picks it.

Creative is the new targeting. Delivery decides who sees an ad, which leaves what the ad says and shows as the lever you control. The creative is the argument.

That last job separates agencies more than any tactic. An account can lift in platform ROAS and shrink the business, because it judges an ad by the transactions credited inside an attribution window, and Meta documents how that credit is assigned in its attribution system help article. Buy discount hunters who never come back and the dashboard still looks fine. We judge acquisition on lifetime gross profit to CAC: the profit a customer produces across their life against what it cost to acquire them. CAC is customer acquisition cost, and the math is in eCommerce unit economics.

What Hayes Media does

The scope, in roughly the order we build it.

• Audit the account and the creative library first. Every ad that has run, what it spent, what it returned. Which angles earned and which formats died is the cheapest research available, and nobody reads it.
• Set the scoreboard before we set budgets. Lifetime gross profit to CAC, with in platform numbers as a supporting view. We agree where that figure comes from and who pulls it.
• Build the creative strategy. Angles, offers, objections, and customer beliefs, mapped to a production plan rather than a vibe. The whole flywheel runs on this document.
• Produce direct response creative. Our homepage states the standard the work is held to: "Our Performance Creative Process has produced winning ads that have spent $100,000's with on-target metrics."
• Run a testing cadence with a kill rule. Every concept goes live against a defined hit rate. Our heuristic is a hook rate above 40% before a concept earns more production time, which is our bar, not a platform standard.
• Build iteration trees off the winners. A concept that works gets rebuilt: new hooks on the same body, new formats, new creators reading the same argument. Most of a mature account's spend should sit on descendants of proven ads, not on fresh guesses.
• Run the media buying. Our homepage puts it plainly: "We run your ads using the media buying strategies that we've refined over 8 years and millions in ad spend." Structure, budgets, and scaling calls, with exclusions set so paid social isn't paying again for customers you already have.
• Keep the creative flywheel turning. Production feeds testing, testing feeds iteration, iteration feeds production. The flywheel is the product. Winners are the output.
• Take email and SMS from the same building if you want them. Then what a customer costs and what a customer is worth get decided by one team. See done for you eCommerce marketing.

Three commercial terms worth stating up front: no onboarding fees, no lock in contracts, no junior marketers. The work sits on our ad creative portfolio, described there as "A slice of the static creative we build for eCommerce brands on paid social."

Who this is for, and who it is not for

This is for you if:

• You're doing at least $100K per month in revenue and want to scale past $10M a year. The qualifier is the $100K a month. The outcome is the $10M, which is the aim stated on hayesmedia.co.
• Your creative pipeline is the bottleneck. You have budget you can't deploy because you don't have enough good ads to put behind it. The most common stuck account.
• Your CAC has climbed and nobody can tell you why. Usually the answer is a library that stopped producing winners and an account living off two fatiguing ads.
• You want the program judged on profit, not a dashboard. Lifetime gross profit to CAC is the number we report and the number we want to be held to.

This is not for you if:

• You're pre revenue or pre product market fit. Build organic proof first. Paid social faithfully scales a weak offer into a larger loss.
• You want a media buyer only. Buying without creative is a better driver for a car with no fuel. A freelance buyer is fine, it just isn't this.
• You want to approve every ad personally. Volume is the mechanism, and approvals cap the flywheel at whatever the founder has time to review.
• You want the cheapest option. We ranked the field, including shops with lower entry points than ours, at best Meta ads agencies for eCommerce.
• You already have a senior in house buyer and a working studio. Then yours is a staffing question, and we wrote the honest version at eCommerce agency vs in house.

What it costs to get Meta ads wrong

No invented figures here. The damage shows up as mechanism, not a tidy percentage.

Creative starvation. An account running four ads has four chances to win. When those fatigue, the only response left is bidding harder behind tired creative. Spend rises, returns don't.

Optimizing to the wrong scoreboard. An agency graded on in platform ROAS finds the cheapest transactions available, usually discount driven purchases from people already close to buying. Reported ROAS rises. Gross profit per customer falls. Both are true at once, and that argument never resolves without a profit view.

Testing without a kill rule. Concepts that should have been stopped keep spending because nobody set the threshold in advance. The account fills with mediocre ads, each defensible and collectively expensive.

No iteration off winners. A brand finds a winning ad, celebrates, then goes hunting for a completely new concept. The cheapest performance available is usually the fifteenth variation of the ad that already worked.

Counting repeat orders as new customers. Paid social takes credit for people who were already yours, so reported CAC reads lower than the real cost of adding a customer. New customer CAC and blended CAC answer different questions.

How an engagement runs

The sequence is the same on every account, and we never skip the first step, because everything after it depends on what it finds.

1. Audit the account and the library. Spend history, structure, exclusions, and every ad that has run with what it earned. That tells us which arguments your customers already responded to.
2. Agree the scoreboard. Lifetime gross profit to CAC, with the data source named and the report format agreed before anything is built.
3. Write the creative strategy. Angles, offers, and objections mapped to concepts, with the production plan attached.
4. Clean up the account structure. Consolidation, exclusions, and budget logic, so the testing that follows produces readable signal instead of noise spread across fragmented ad sets.
5. Ship the first production batch. Concepts go live against the defined hit rate and kill rule.
6. Read the results and cut. Winners get flagged for iteration, losers get stopped, nothing survives on sentiment.
7. Build iteration trees. New hooks, formats, and creators on the concepts that earned it, until the tree stops producing.
8. Scale behind proven creative. Budget goes up behind ads that have already earned a hit rate, not behind hope.
9. Report, then repeat. What spent, what it returned, what those customers are worth, and what changes next.

The proof

Every figure below is quoted exactly as it appears on our own pages.

From hayesmedia.co:

• "Trusted by 57+ eCom brands"
• "Meta ad creative, Meta media buying, Email & SMS. Take all three or just the one you need, and scale past $10M/year."
• "Our Performance Creative Process has produced winning ads that have spent $100,000's with on-target metrics."
• "We run your ads using the media buying strategies that we've refined over 8 years and millions in ad spend."

Two acquisition case studies, quoted as published:

• Remi. Figures as published: "12,400% Revenue Growth" / "150% ROAS Increase".
• Mammoth Headwear. Figures as published: "419% Revenue Growth".

One line in the Remi study isn't a percentage. The work "turned what was an unprofitable CAC to a profitable CAC for new customer acquisition, and led to tremendous revenue and LTV growth." A revenue percentage says the account got bigger. A CAC crossing from unprofitable to profitable says the business got better.

What those founders say, from the same pages:

• "One of the biggest differences between working with Hayes, and working with the other agencies in the past, is that the Hayes team are just so proactive." Oscar Adelman, Founder & CEO of Remi
• "They've done a great job keeping the ROAS very profitable, which has been amazing for my company. Jordan and the team really go above and beyond." Taylor Holst, Founder & CEO of Mammoth Headwear

One honest limit, the same one we gave ourselves in our ranked list: those case studies are published by us. We have no third party review profile.

How to choose a Meta ads agency

Run this on us and on everyone else you're shortlisting. It takes one call.

1. Ask what number they'll be judged on. If the answer is in platform ROAS and nothing else, ask what stops them from buying discount hunters to hit it. A good answer names gross profit or contribution.
2. Ask how many creatives they'll test per month, and where those come from. In house studio, contracted editors, or your own footage recycled. That tells you whether they're a creative shop with buying attached or a buying shop hoping you supply the ads.
3. Ask what their hit rate is. Every honest operator knows roughly what fraction of their concepts become winners. An agency that has never measured its win rate has never been held to one.
4. Ask what happens to a miss, and to a winner. A miss should die against a threshold set in advance, not by argument. A winner should be rebuilt: new hooks, formats, and creators off the proven argument. If the answer for a winner is just scale it, half the job is missing.
5. Ask who touches the account. Names and seniority, not a team page. Then ask how many accounts that person carries.
6. Ask whether they'll touch the landing page and the offer. An agency that treats everything after the click as your problem drew that line around their convenience, not your revenue.
7. Ask what leaving looks like. Ad account ownership, creative files, raw footage, and data. This is the one brands most regret not asking.
8. Ask what they won't do. An agency that serves everyone at every size and in every category has not thought about fit.