Facebook ads agency vs in house: what each really costs, what each produces, and who should do which

If you spend less than about 50,000 a month on Meta and your ads are not being refreshed every week, an agency is usually the better buy, because the hard part of Meta in 2026 is making enough good new ads, and one in-house hire cannot do that alone. In-house wins when you already have a creative team producing ads every week and need someone close to the business to run the account day to day. Most brands that spend seriously end up with a split: strategy and approvals stay inside, and ad production and testing volume come from outside.

Most of the pages ranking for this search compare a media buyer's salary with an agency's monthly fee and stop there. That made sense when the job was mostly setting up audiences and moving budgets. It misses where the money and the results now sit, which is the ads themselves. This piece compares the two on cost, on output, and on speed, using what we see across the accounts we run.

What you are really choosing between

The question used to be "who presses the buttons in Ads Manager". With broad targeting and Meta's own systems deciding who sees what, the buttons matter less every year. Meta describes its ad retrieval system as built to match each person with the most relevant ads out of a very large and growing pool of creative (Meta Engineering). In plain terms, the ad now does most of the targeting. A different idea finds a different buyer.

So the real choice is between two ways of getting three jobs done:

1. Deciding what to say. Which buyer, which moment, which argument. This is strategy.
2. Making the ads. Statics and videos, in the right sizes, with copy, every week.
3. Running and reading the account. Launching, judging what worked against your target, and moving spend.

An in-house hire is usually strong on the third job, reasonable on the first, and overloaded on the second. A good agency should cover all three, but it will always know your product less well than you do. The rest of this article is about where that trade lands for you.

What we see in our own accounts backs this up. Over the last 90 days we logged 610 lessons across 15 accounts about what worked and what did not. Of the lessons that judged an idea against the account's target, 148 cleared it and 335 missed. That is roughly one in three. Nobody, in-house or agency, picks the winners in advance. The team that ships the most genuinely different ideas finds them first.

The cost, side by side

Here is the comparison most guides make, with the line they leave out added at the bottom.

	In-house	Agency
People cost	A digital marketing manager from about $72,000 a year before benefits (Databox, citing Glassdoor)	About $2,000 a month on average, up to around $10,000 for premium agencies (Databox), or 10 to 30 percent of spend on a percentage model (SaaS Hero)
Time to useful results	3 to 6 months just to recruit, by one agency's estimate, before any ramp up (SaaS Hero)	Weeks, if they have run accounts like yours
Knowledge of your product	Deep, and it grows	Has to be built, and has to be kept up to date by you
Creative production	Not included. Needs designers, editors, creators or a studio on top	Depends entirely on the agency. Ask how many new ads a week, in writing
Cover when someone leaves	Starts again	Continues

That last line about creative is the one that decides most of these choices. A salary buys you a person. It does not buy you 20 new ads a week. If your in-house buyer has to brief, chase and approve every ad from freelancers, most of their week goes on production, and the account gets two or three new ads a week instead of the volume your budget can read. We worked through how many ads a given budget can actually test in how many ads to run on Meta, and for most brands spending 10,000 to 100,000 a month the answer is far more than one person can make.

The hidden cost on the agency side is the opposite problem. Some agencies are mostly media buying with a few ads a month attached. If the fee is a percentage of spend, their income grows when you spend more, whether or not the ads got better. Neither model is cheap if the ads stop changing.

What each can actually produce in a month

This is the gap the other guides miss, so it is worth being specific.

We built more than 600 new ads in the last 60 days. We make twenty to fifty new ads a week for each client because, as the table above shows, roughly two in three ideas miss and you cannot tell which in advance. A single in-house marketer who is also running the account, reporting, and answering the founder's questions on a Monday cannot match that, however good they are. They need a production team, and once you have built one you have built a small agency inside your business.

Volume on its own is not the answer either. Five versions of the same video with new first lines are one idea, not five. What moves accounts is variety in who the ad is for and what it says. Two things we saw this year make the point:

• The buyer on the fence beat the fan. On a UK luxury home brand, the ad written for the person who hesitates on price was the clear winner at nearly double the target return, while ads written for people who already loved the category missed. We wrote it up in the first August Weekly. An in-house team that knows its loyal customers well tends to write for them. That knowledge can be a blind spot.
• Proof beat the guarantee. On a UK luxury laundry brand, ads built on social proof came in about a fifth over the return target, while ads built on a money-back promise took more of the budget and returned about half of target (August Weekly, 9 October). Nobody inside the business had a reason to doubt the guarantee. Testing both side by side settled it.

Whoever you choose, the useful question is not "agency or in-house" but "who is going to put five to ten genuinely different ideas in front of buyers every week, and who is going to judge them honestly".

Who should do which

A rough guide, by where your business is now.

Choose an agency if:

• You spend roughly 10,000 to 100,000 a month and have no creative team.
• Your best ads are months old and nothing new has beaten them.
• You need results this quarter, not after a hiring process.
• You want people who have seen the same problem on other accounts this month.

Choose in-house if:

• You already produce new ads every week from your own team or creators.
• Your product needs deep explaining, regulated claims or constant sign-off, and the person running ads must sit next to whoever approves them.
• Your spend is large and steady enough to justify a full team, not one hire..

Choose a split if:

• You have someone good inside who knows the customer, but they are buried in production.
• Keep the customer, the offer and final approval in-house. Bring in outside help for ad volume and testing, with clear weekly numbers.

The split is where most of the brands we work with land. The founder or marketing lead owns the brand and says yes or no. We bring the ads, the testing and the read on what worked. Neither side tries to do the other's job.

How to judge an agency before you sign

If you go outside, these questions separate the agencies that make ads from those that mainly manage them.

• How many new ads will go live in my account each week? Get a number. "As many as needed" is not one.
• Who writes them, and from what? The answer should start from your customers, your site and your account data, not a template.
• How do you decide an ad has won or lost? Look for a target, a minimum spend before a verdict, and a willingness to say an ad was too early to call. In one account this autumn an ad that looked like the scale candidate one week fell apart the next, with its hook rate and click rate still healthy. It was a conversion problem, not tired creative, and the fix was different. You want a team that tells those apart.
• What do you do when an ad gets tired? We flagged 19 ads as fatigued across our accounts in the last 90 days. The right response was a fresh version of the same winning idea, not a new offer. Ask what theirs is.
• Can I see results you are allowed to show? Ours are on the case studies page, named, with the timeframes.

Red flags: a minimum term before you have seen a single ad, no weekly creative number, reporting that only shows platform ROAS with no link to your store's orders, and nobody able to tell you why last week's winner won.

How to make in-house work if you choose it

In-house can work very well. When it does, it is because the business treated ad production as a job in its own right.

• Hire for judgement on the account, then fund production separately: a designer or editor, a creator budget, or an outside studio.
• Set a weekly number of new ads, tied to your testing budget. Our guide to creative testing budgets shows the maths.
• Write every ad for one buyer at one moment. Most of the losers in our data were ads that tried to speak to everyone.
• Keep a written log of what won and what lost, with the reason. It is the asset that makes month six better than month one, and it leaves with the person if it lives only in their head.

FAQ

Is a Facebook ads agency cheaper than hiring in-house? Usually, yes, for the people cost alone. Average agency fees sit well below a full-time salary (Databox). The bigger difference is creative: an in-house hire needs a production budget on top, while a good agency should include new ads every week.

When should a brand move Meta ads in-house? When you already produce new ads every week and your spend supports a team rather than one person. Moving in-house to save a fee, without a production plan, usually means fewer new ads and a slower account.

Can I use an agency and an in-house marketer together? Yes, and for many brands it is the best setup. Keep the customer, the offer and approvals inside. Use the agency for ad volume, testing and the weekly read.

How quickly should an agency get new ads live? Within the first week or two. We start new clients with 30 ads in 72 hours, and you can see what that involves on our sprint page.