A single country-wide Meta ad set gives you one cost per purchase, blended across every state it spends in. If your store's order data shows a few states drive most of your sales, give those states their own ad set and most of the budget, target the rest of the country with them excluded, and keep low-converting areas on a small budget. We usually push consolidation on Meta. This is a split we'll argue for, because the ad sets target different places and the split comes from your customer data, not a hunch.
The account: one setting for the whole country
On a mentoring call, we spoke with an in-house marketer at an ecommerce brand selling an $80 product across the US. Meta took about 90% of the spend, around $20,000 a week, and the business was ready to spend far more. The target cost per acquisition was $50. Pushing spend up kept dragging it to $64 or $65. The location setting was simply United States.
We asked which states the most orders came from, by delivery or billing address. The answer came straight back: California, Texas and Florida.
Our advice was to build an ad set for just those three states and see what the CAC is there, then divide the rest of the country by what the order data says. The areas producing the most orders get the most budget. Areas that don't convert get a minimal budget, kept running for brand presence. Our read on the call was that a national ad set also spends in expensive metro areas that may convert worse, and that pulls the blended number up. That's a hypothesis until you check it.
Check before you split
- Pull orders by state from your store, by shipping or billing address.
- Pull Meta's region breakdown for the same period. Meta's breakdowns documentation lists country, region and DMA, bases region on information such as a person's hometown and current city, and says breakdown values are estimated. Treat it as directional.
- Compare share of spend with share of orders. States taking far more of the spend than they return in orders are your case for splitting. If the two line up, one ad set is doing its job.
Then build it. Meta's targeting reference lets an ad set target regions (a state, province or region), cities with a radius, or ZIP codes, and exclude locations. One ad set targets your top states. A second targets the country with those states excluded.
When a geographic split is justified
This doesn't contradict why one Meta campaign beats five. That problem is ad sets aimed at overlapping audiences, bidding in the same auctions and splitting the same conversions. A state split with exclusions points each ad set at a different place. It's justified when:
- Your orders clearly cluster. The split follows your data, not a guess about which cities feel valuable.
- Each ad set can still learn. At $20,000 a week that's realistic. On a small budget, splitting can leave both ad sets short of the conversions they need to get through the learning phase, so keep one.
- You keep it to a few ad sets. Top states and everywhere else, maybe a middle tier. Not one per state.
It's the same logic we apply on Google when one campaign covers the US and Australia: when one blended average hides very different markets, separate them so you can fund the one that pays.
The takeaways
- A national Meta ad set reports one blended CAC across every state it spends in.
- Compare your orders by state with Meta's region breakdown before you change anything.
- If orders cluster, give the top states their own ad set and most of the budget, and exclude them from a second ad set for the rest.
- Only split when each ad set has enough budget to learn. Small budgets stay in one ad set.
Common questions
Should I split my Meta ad set by state?
When your own order data shows that a few states drive most of your sales, and your budget is large enough for each ad set to get regular conversions, yes, it is worth testing. Give the top states their own ad set and most of the budget, and target the rest of the country with those states excluded. With a small budget, keep one ad set.
How do I see Meta ad results by state?
Use the region breakdown in Meta's reporting. Meta's Marketing API documentation lists country, region and DMA as breakdowns, says region is based on information such as a person's hometown and current city, and notes that breakdown values are estimated. Compare it with orders by shipping or billing state from your store.
Can Meta target or exclude specific states?
Yes. Meta's targeting documentation lets an ad set target regions, described as a state, province or region, as well as countries, cities with a radius, ZIP codes and custom locations. It also supports excluded locations, so one ad set can target your top states and another the rest of the country without them.
Doesn't splitting ad sets go against consolidating on Meta?
Consolidation fixes ad sets that chase the same people and split the same conversions. A state split with exclusions aims each ad set at a different place, and it is based on where your customers are. It still needs enough budget per ad set to learn, which is why we keep it to a few ad sets, not one per state.
How much budget should low-converting states get?
Our approach is to put most of the budget on the states that produce the most orders and keep a small budget running in the rest, for brand presence, rather than switching them off. Revisit the split as your order data changes.