Meta Ads Agency.
We give the algorithm
something worth reading.

Two costs, and only one of them is the agency. Your ad budget goes to Meta. The management fee goes to whoever runs the account.

Published market rates for management sit at roughly 10 to 20 percent of monthly ad spend, or a flat monthly retainer, usually with a one time setup fee covering the tracking and campaign build. Creative production is normally billed on top of all of it. We quote a figure instead of a formula. The setup fee is charged once in month one, the management fee runs from launch, and from month two the management fee is the only thing that recurs.

The fee is quoted against your account and the spend level it covers, so a larger budget is quoted at a larger figure. What it will not do is recalculate itself every month. It is agreed in writing before anything starts, and it holds until a new figure is quoted and approved by you.

Deciding what the next ad should say is included. Filming it is not.

Writing the copy, choosing the angles, directing the concept and cutting footage you supply into ad variants all sit inside the management fee. Producing raw video or photography does not, and it gets quoted separately in writing before anyone starts. That is the normal split across the industry, where graphics and video are billed per asset on top of a retainer.

Plenty of clients shoot their own footage on a phone and hand it over. On Meta that often outperforms studio work, and it costs nothing extra.

Enough to buy conversions the system can learn from. That depends far more on your cost per result than on any round number.

A business paying twenty dollars a lead and one paying four hundred dollars a sale need very different budgets to produce the same amount of learning in the same window. So the question worth asking is how many conversions a month the budget can realistically produce, not how many dollars it is.

If the honest answer is a handful, the budget is too thin to test creative properly. We would rather say that before taking the work than after.

Not with us. There is nothing long to sign and nothing taken in advance.

What does get agreed at the start is a realistic working horizon, because a creative testing cycle cancelled after a few weeks has cost you money and taught you nothing. That is a planning conversation rather than a lock-in clause.

Scope and fee are written down before any budget moves. If it stops being worth doing, it stops.

Campaign strategy, structure and objectives across Facebook and Instagram, conversion tracking, audience setup, ongoing optimization, and reporting tied to revenue rather than to reach.

In practice the work has shifted. Conversions API running alongside the Pixel with events deduplicated, Event Match Quality kept high enough to optimize on, and a creative testing schedule that keeps producing new angles are now the parts that decide whether an account works. Audience configuration matters much less than it did.

A Facebook Ads management agency spending its time on interest lists and ad set naming conventions is optimizing for a system Meta has already replaced.

Yes, and it is usually the faster start, because the account already has history worth reading.

Partner access is taken to your existing Business Portfolio rather than rebuilding somewhere new, so the pixel history, the audiences and the accumulated learning all stay where they are. The first thing to look at is the event setup, because that is what everything else has been learning from.

A takeover carries no setup fee. The audit comes first and costs nothing, whether or not anything follows it.

You do. All of it, from the first day.

Your Business Portfolio, ad account, pixel, Conversions API setup, catalog and audience lists belong to you. An agency should be working inside your assets with partner access rather than holding them inside its own, which is what makes a handover a permissions change instead of a negotiation.

If you leave, access is removed and everything stays exactly where it is. Nothing has to be rebuilt and nothing gets held back.

The Meta Pixel reports from the visitor's browser. The Conversions API reports from your server. The difference is what can block them.

Browser events can be intercepted by privacy settings, tracking prevention and ad blockers, so a share of what actually happened never reaches Meta. Server events are sent directly and are not exposed to any of that. Running both and deduplicating them on a shared event ID gives the delivery system the fullest picture without counting anything twice.

The Pixel on its own is no longer enough. It is the single most common thing missing from an account that reports better than it performs.

Event Match Quality is a score Meta gives each conversion event, based on how well the customer information sent with it can be matched to a real person.

It is visible per event inside Events Manager. A low score means Meta is receiving conversions it cannot attribute confidently, so it is optimizing on a partial picture and the cost per result reflects that. Sending more matched fields, securely and consistently, is usually what moves it.

It is worth reading before a campaign spends anything. Optimizing on top of a weak score just teaches the system faster in the wrong direction.

Broad, in almost every account, with your own customer data used as the seed.

Meta's delivery system reads the creative and predicts who it is for. Stacking detailed interests shrinks the pool it is allowed to search, which raises what you pay in the auction without improving who arrives. Advantage+ Audience given a real customer list consistently finds buyers a hand-built interest set would have excluded.

There are exceptions. Regulated categories, narrow professional audiences and small geographic markets still need constraints, and those get set deliberately rather than out of habit.

Long enough for the account to produce conversions the system can learn from. That is a volume question rather than a calendar one, which is why nobody honest gives you a date.

What the early period buys is worth naming. The tracking gets verified, the first creative angles get tested against each other, and the account starts producing data worth acting on. An account being rebuilt often improves sooner than a new one, because fixing the event setup improves signal quality immediately.

You get told what changed and why while that is happening, rather than a report at the end of it.

It depends on whether people are already searching for what you sell.

Where there is real search demand and a clear action worth counting, paid search usually gets there faster and that budget belongs in Google Ads management. Where the product sells by being seen, or the category is one nobody thinks to type, paid social is where the demand has to be created rather than captured.

Most accounts end up running both. Which one goes first is a question about your market, and it deserves an honest answer rather than whichever page you happened to land on.

Yes, though not through this page.

Agency partnerships run under a separate arrangement with its own reporting, its own communication rules and its own commercial terms, because the sell price, the promised deliverables and the margin all belong to you rather than to us. That is set out on our white label PPC page.

This page is written for businesses running ads for themselves.