Meta says 40 sales, my store says 26. Who is lying?

Neither. They are counting two different things. Meta credits view-through conversions where someone saw an ad, did not click, and later arrived on their own; its default window counts a click within seven days; people browse on a phone and buy on a laptop; and where the pixel and the Conversions API both fire without deduplication the same sale is reported twice. Our answer is to report marketing efficiency ratio — total revenue divided by total ad spend — as the number that decides things, and use Meta's figures only to compare one ad against another.

Why do you run so few ad sets?

Because each one needs roughly 50 conversions a week to leave the learning phase and behave predictably. Splitting a small daily budget across ten ad sets funds each of them at a level where nothing is ever learned, and then the account is blamed for a structural decision. One ad set with enough budget to exit learning beats ten that are permanently in the dark.

How much creative do you need from us?

If creative production is not part of the engagement, expect to supply three to five new ads a week. That sounds like a lot until you see how fast creative fatigues here. What matters is angles rather than variations — a different reason to care, not the same image with a new headline. We will write the brief and tell you exactly what to shoot, and where we are producing the creative it is a separate line in the proposal because it is genuinely separate work.

Does tracking still work after the Apple changes?

Partly, and it is worth being honest about it. The 2021 iOS prompt removed a large share of browser-based signal and it has not come back. Server-side reporting through the Conversions API recovers a meaningful part of it and improves what the algorithm can learn from, which is why we set it up before spending. But anyone telling you the reporting is as complete as it was in 2019 has not looked.

Can you just run the account if we handle creative ourselves?

Yes, and it is a lower rate. It also puts a large part of the outcome outside our control, so we write that into the arrangement rather than leave it unsaid. On Meta the creative is the campaign — targeting has become progressively more automated, and what still separates a working account from a failing one is the first second of the ad.

How long before we know if this works?

Plan on the first month being partly an investment in finding out. The learning phase alone takes a couple of weeks, and a fair read on whether the channel suits your product and margin usually needs about 45 days. What you get in the meantime is a weekly account of what has been tested and what it told us.

Paid Media & Performance

Meta ads for Facebook and Instagram

Fewer ad sets with enough budget to learn, a weekly supply of new creative angles, and a monthly number that reconciles with your bank rather than with Meta.

From: $300–$2,500/mo
Typical time: ongoing
Usually bought by: DTC, local services

What this actually solves

Nobody on Facebook or Instagram is looking for you. That single fact decides everything about how this channel is run. On Google, somebody types a problem and you answer it. Here, somebody is scrolling through pictures of friends and holidays and you have roughly one second to be more interesting than the next thumb movement. Which is why the entire craft sits in the creative — the first frame, the first line, the reason a stranger should care — and why an account can be structured immaculately and still fail.

Two things bring buyers here. The first is reach at a price Google cannot match; cost per thousand impressions on Meta is usually well below Search, so it remains the cheapest way to put a new product in front of a large number of people. The second is that Meta reaches people who do not yet know they want the thing. On Search you capture someone who has already decided to remodel a bathroom. Here you reach the person who had not considered it until a good video made them consider it. For direct-to-consumer brands and local service businesses, that is where growth beyond existing demand actually comes from.

There is a caveat we would rather state before you buy than after. Apple's 2021 tracking prompt removed a large part of the browser-based signal this platform ran on, and the brands it hurt most were the small ones with nothing else holding acquisition up. Server-side reporting and stronger creative make the channel work again, but the reporting is not as complete as it once was, and any monthly report that treats Meta's own conversion count as the truth will overstate what happened.

What we do, and in what order

Get the measurement right before spending anything. Pixel and Conversions API both installed, events fired and tested end to end, and deduplication verified so a single purchase is not counted twice. We also reconcile Meta's reported sales against the store's own figures on day one, so we know the size of the gap before it becomes an argument. Two to four days.
Build a structure that can be read at a glance. Typically three campaigns — people who have never heard of you, people who have engaged or visited, and your existing customers — with deliberately few ad sets, each funded well enough to reach the roughly 50 conversions a week that stabilise delivery. One to two days.
Set up the creative system, not a set of ads. Angles first — the underlying reason someone would care — then the executions. Hooks written for the first two to three seconds, primary text kept under about 125 characters so it is not cut off behind See more, headlines short. Enough to launch with genuine variety rather than one idea in four colours. Three to five days for the first batch.
Launch, then watch daily and change weekly. Looking at a campaign does not disturb the learning phase; editing it does. The first fortnight is checked every morning for budget errors, disapprovals and broken landing pages, while changes are batched into a weekly pass. Two weeks, then ongoing.
Weekly creative rotation and a monthly report that reconciles. Fatigued ads retired, winners iterated into new versions, frequency watched, and a monthly report built on marketing efficiency ratio rather than platform-reported return. Ongoing.

Tools used — Meta Ads Manager, Events Manager and Business Manager, the Meta Ads Library for competitive reference, Canva and CapCut for production, and Foreplay or Motion where the creative volume justifies a proper library.

What moves the price

The engagement is $300–$2,500/mo and ongoing.

Whether creative production is included. This is the biggest single factor. Running an account where you supply the assets is one rate; producing three to five new ads a week is a different service with its own cost, and we itemise it rather than fold it in and then quietly under-deliver on it.

The size of the ad budget. A larger budget means more consequence attached to every decision, more frequent checking, and more scrutiny on reconciliation between platform figures and actual revenue.

How many products, markets or offers are in play. One offer to one country is a straightforward account. Several product lines, seasonal offers and multiple markets multiply the creative requirement and the number of things that can quietly break.

What usually goes wrong

The budget is split across too many ad sets. Ten ad sets on twenty dollars a day gives each of them two dollars — far below what any of them needs to leave the learning phase. The account then produces noise rather than data, every result looks random, and the conclusion drawn is that Meta does not work. Consolidating into fewer, better-funded ad sets is often the entire fix.

Creative is treated as a launch task rather than a weekly supply. The same three ads run for two months, frequency climbs past three, results decay, and budget gets increased to compensate — which only shows the same tired ad to the same people more often. On this platform creative fatigue is not an edge case, it is the normal state, and the plan has to assume it.

The purchase event is counted twice. A store running the platform's native integration while a second pixel sits in Tag Manager will report every sale twice. Return on ad spend reads four when it is two, you are delighted, and the bank balance disagrees. The check takes ten minutes — compare the platform's sales count to the store's for the same period, and if it is close to double, the cause is in the setup rather than the market.

See where the site stands before you brief anyone.

If the site is already live, start from what it scores rather than from a guess. The audit reports what it checked, what it found, and — the part most tools leave out — what it could not check, and why.

See the website audit Talk about this work

Questions people ask before they commit.

Meta says 40 sales, my store says 26. Who is lying?: Neither. They are counting two different things. Meta credits view-through conversions where someone saw an ad, did not click, and later arrived on their own; its default window counts a click within seven days; people browse on a phone and buy on a laptop; and where the pixel and the Conversions API both fire without deduplication the same sale is reported twice. Our answer is to report marketing efficiency ratio — total revenue divided by total ad spend — as the number that decides things, and use Meta's figures only to compare one ad against another.
Why do you run so few ad sets?: Because each one needs roughly 50 conversions a week to leave the learning phase and behave predictably. Splitting a small daily budget across ten ad sets funds each of them at a level where nothing is ever learned, and then the account is blamed for a structural decision. One ad set with enough budget to exit learning beats ten that are permanently in the dark.
How much creative do you need from us?: If creative production is not part of the engagement, expect to supply three to five new ads a week. That sounds like a lot until you see how fast creative fatigues here. What matters is angles rather than variations — a different reason to care, not the same image with a new headline. We will write the brief and tell you exactly what to shoot, and where we are producing the creative it is a separate line in the proposal because it is genuinely separate work.
Does tracking still work after the Apple changes?: Partly, and it is worth being honest about it. The 2021 iOS prompt removed a large share of browser-based signal and it has not come back. Server-side reporting through the Conversions API recovers a meaningful part of it and improves what the algorithm can learn from, which is why we set it up before spending. But anyone telling you the reporting is as complete as it was in 2019 has not looked.
Can you just run the account if we handle creative ourselves?: Yes, and it is a lower rate. It also puts a large part of the outcome outside our control, so we write that into the arrangement rather than leave it unsaid. On Meta the creative is the campaign — targeting has become progressively more automated, and what still separates a working account from a failing one is the first second of the ad.
How long before we know if this works?: Plan on the first month being partly an investment in finding out. The learning phase alone takes a couple of weeks, and a fair read on whether the channel suits your product and margin usually needs about 45 days. What you get in the meantime is a weekly account of what has been tested and what it told us.

Often taken together.

Start with the brief, not the invoice.

Send what you are trying to achieve. You get a written scope, a price and a date — or an honest “this is not the right service for that”.

Talk about Meta (Facebook / Instagram) ads All Paid Media & Performance services