For a while, Advantage+ sales campaigns came with a small, sensible control. You could tell Meta what share of your budget it was allowed to spend on people who had already bought from you. Ten percent, twenty, zero. The campaign did the rest.
That control is gone. Meta's own help page now says the existing customer budget cap "is no longer available," and points you to a manual workaround instead. The campaign did not lose interest in your customers when the setting disappeared. If anything, it can now chase them with nothing in the way.
This matters to more businesses than the name suggests. Advantage+ sales is not just for online stores. Plenty of salons, clinics, studios and home service companies run the Sales objective because it optimizes toward a booking or a purchase on their website. If that is you, some part of your budget is probably being spent on people who were coming back anyway, and your dashboard is calling it a win.
Here is how that happens, and how to put the control back.
Meta Retired the Existing Customer Cap. The Campaign Still Chases Your Customers.
Advantage+ sales is Meta's most automated campaign type. You hand it a budget, a goal and your creative, and it decides who sees the ads, where and when. (Whether to hand Meta that much control in the first place is its own question, and we worked through it in Should You Let Meta Run Your Whole Campaign?) The old cap was one of the few levers you kept. It drew a line around your existing customers and said: this much, and no more.
Without that line, the campaign does what any optimizer does. It goes where results come easiest. And nobody converts more easily than someone who already knows your name, already trusts your work, and already has your number saved in their phone.
A past customer is the cheapest result Meta can find. That is exactly why it keeps finding them.
This is not a malfunction. From the system's point of view, a purchase is a purchase. It has no way of knowing that a woman who books her third haircut with you this year would have booked it from your reminder text, or that the man reordering the same filters every quarter was going to reorder them regardless. It sees a conversion, credits the ad, and spends more in that direction.
Meta is not hiding this, to its credit. In its own short video on Advantage+ sales, Meta recommends excluding existing customers when "your primary goal is net new customer acquisition." The company knows the campaign leans toward people who have already bought. It simply stopped giving you a one-click way to manage it.
A Sale From a Past Customer Looks Exactly Like a New One on Your Dashboard
The reason this goes unnoticed for months is that the numbers look fantastic.
Spend on existing customers almost always reports a better return than spend on strangers. Past buyers click more, buy faster and need less convincing. So when a larger share of the budget drifts their way, cost per purchase drops, return on ad spend climbs, and the account looks like it is getting better at its job.
Two things are wrong with that picture.
The first is attribution. Meta's default setting counts a purchase if someone clicked an ad within the past seven days or saw one the day before. A regular customer sees your ad on Tuesday, gets your newsletter on Wednesday, and books on Thursday. Meta counts that sale. So does your email platform. You made one sale and two systems are claiming it. We covered how far those claims can drift from your bank balance in Meta Says It Made You 40 Sales. Your Bank Says 12., and why the attribution setting shapes who Meta goes looking for in That Attribution Setting Is Not a Report. It Is an Instruction.
The second problem is quieter: diminishing returns. The first few dollars you spend reminding customers you exist can genuinely bring people back sooner. Every dollar after that buys less, because you are paying to reach the same people again and again.
Picture a boutique fitness studio spending $600 a month on Meta, with about half of it drifting to current and lapsed members. The campaign reports 30 bookings from that group. Now suppose the studio cut that half to $300. How many of those 30 disappear? If the studio also sends a weekly email and a class reminder by text, the honest answer is probably very few. Most of those members were booking anyway. The extra $300 bought a handful of bookings at best, and the dashboard spread the credit for all 30 across the whole spend.
You cannot see this in a blended number. A return figure averaged across new and existing customers hides the fact that one half of the spend is doing the hard work of finding strangers, and the other half is taking credit for your regulars.
Meta Can't Cap Customers You Never Uploaded
Before any control works, Meta has to know who your customers are. This is the step most accounts skip, and every fix below depends on it.
Start with a customer list. Export names, emails and phone numbers from your booking system, your point-of-sale, your CRM or your store platform, and upload them as a customer list custom audience. Meta matches those details against its own accounts. It will not match everyone, because people use different emails and numbers on Facebook than the ones they give you, so the more identifiers you include per person, the more of your list Meta can recognize. Meta's formatting guidelines suggest starting with at least 1,000 customers. A smaller business can still upload what it has; it just means your website audience has to carry more of the load.
That second audience is a purchaser audience built from your website: everyone who fired a purchase or booking event in the last 180 days. It catches customers whose details never matched, and it updates itself.
Then tell the ad account what those audiences mean. In your advertising settings, Meta lets you define an existing customers audience and an engaged audience (people who know you but have not bought). Once those are set, your sales campaign reports can be broken down by audience segment: new audience, engaged audience and existing customers. That breakdown is how you will judge everything that follows, so set it up even if you change nothing else.
One practical note: a customer list is a snapshot. Anyone who buys next month is not on it until you upload again. Refresh it at least monthly, or connect your booking or store platform so it syncs on its own.
Excluding Every Past Customer Is the Wrong Fix. A Spending Limit Is the Right One.
The obvious fix is to exclude every existing customer from the campaign. Usually, that is a mistake.
We made that argument at length in Your Cheapest Meta Customers Are the Ones Who Already Know You. Past buyers are often your best prospects for a second purchase, keeping them in front of your brand is how you earn referrals, and cutting them out can starve an ad that was about to do its job. None of that changed when Meta removed the cap.
What changed is that you now have to set the limit yourself. The goal is not zero spend on existing customers. It is a deliberate, small share instead of whatever the campaign happens to drift toward.
Meta's help page describes two ways to rebuild the control by hand.
Option one: exclude them entirely. Create the campaign, open the audience section, and under Controls add your existing customer audiences as an exclusion. Every dollar then goes to people who have not bought from you. Use this only when the offer itself is for newcomers, such as a first-visit discount you do not want your regulars to see, or a new-patient special.
Option two: split the campaign in two. This is the one most businesses want, and it is the old cap rebuilt from parts.
- Create one campaign with a campaign-level budget.
- In the first ad set, open Budget and schedule and set an ad set spending limit. Choose a daily or lifetime maximum, either as a percentage of the campaign budget or as a dollar figure. A percentage scales with the budget; a dollar figure stays fixed.
- In that same ad set, switch the audience setup so it includes only your existing customer audiences, and uncheck Use as a suggestion. That checkbox matters. Left on, Meta treats your audience as a hint and is free to wander past it.
- Add your creative.
- Duplicate the ad set. In the copy, remove the included audiences and instead add your existing customers as an exclusion. Keep the same creative in both so the only difference is who each ad set can reach.
Meta will aim to stay under the maximum and will often spend less than it. That is fine. The limit is a ceiling, not a target.
Spending limits only exist on campaigns with a campaign-level budget, which is one reason the budget decision matters more than it looks. We explain that trade-off in Who Should Hold the Budget, the Campaign or the Ad Set?, and the quirks of how Meta honors spending limits in Meta Won't Give Your New Ad a Fair Shot. Here's How to Force One. This is also the kind of structure our Facebook ads management team rebuilds most often when an account's results look better than its bank balance.
Reaching Every Customer You Have Costs Less Than You Are Spending on Them
So how large should the existing customer share be? Most owners guess, and most guess high. There is a better way: work out what it would cost to reach every customer you have, as often as you want, and treat that as the ceiling.
The arithmetic is simple. Take the number of customers Meta can actually reach, multiply by how many times a month you want each of them to see an ad, divide by 1,000, and multiply by what you pay per thousand impressions (your CPM, which is in your Ads Manager columns).
Say a physical therapy clinic has 2,000 past patients, and 1,200 of them match to accounts Meta can reach. The clinic wants each to see an ad about four times a month. That is 4,800 impressions. At an example CPM of $15, reaching every matched patient four times a month costs about $72.
Seventy-two dollars. Now open the audience segment breakdown and look at what the campaign is actually spending on existing customers. If the answer is $400 a month, you are not buying reach. You are buying the same people's attention five times over.
The right share also depends on how your customers buy.
Businesses built on repeat visits can justify a real, if modest, allocation. A salon, a dental office working through its recall list, a pet groomer, a café with a loyalty crowd, a store selling something people run out of: a timely reminder can pull a visit forward or win back someone drifting toward a competitor. Here the cap is about restraint, not elimination.
Businesses built on rare, big purchases should keep it close to zero. Someone who replaced their roof last spring, bought a hot tub, or had a kitchen remodeled is not a customer to re-acquire. They are a source of referrals and reviews, and those come from how you treated them, not from ads. A roofer spending 30 percent of a sales campaign on last year's customers is paying to remind happy people of a job that is finished.
And for almost every business, repeat purchases come mainly from channels you own: email, text reminders, your own social pages, a note in the box. Those cost next to nothing per message. Paid reach should fill the gaps they leave, not do their job at a higher price.
Cost per New Customer Is the Only Number That Says the Campaign Works
Once the split is running, change the number you watch.
Every week, open the campaign, choose Breakdown, and view results by audience segment. You will see spend, results and cost for new audience, engaged audience and existing customers side by side. The figure that matters is cost per new customer: spend on the new audience segment divided by purchases or bookings from that segment. That is the price of growth. Everything else is maintenance.
Expect the headline numbers to look worse after you cap. Overall cost per purchase will usually rise and reported return will usually fall, because you have taken away the cheapest conversions in the account. That is not the campaign failing. It is the campaign being measured honestly for the first time.
Then check Meta's version against your own records. Your booking system, store platform or CRM knows which purchases came from first-time customers. If Meta says the new audience segment produced 40 purchases last month and your system shows 22 first-time buyers from all sources combined, you have learned something important about how generously Meta counts.
Meta does offer a proper test for this, called a conversion lift test, which holds back a random group of people from seeing your ads and compares what they buy with the group that did. It is the cleanest answer to "would these sales have happened anyway?" It is also out of reach for most small accounts: Meta's guidance is a campaign from the past year with at least $5,000 in spend and at least 500 conversions. Until you are there, the audience breakdown plus your own new-customer count is the honest substitute.
Splitting a Small Campaign in Two Can Starve Both Halves
There is a fair argument against doing any of this, and it deserves a hearing.
Some experienced Meta advertisers argue that carving customers into their own ad sets does more harm than good. Two ad sets split your conversion data in half, and Meta learns faster from one pool than two. A tiny customer ad set may never gather enough results to settle. And in a small account, the overlap between past customers and the people Meta would find anyway may simply not be large enough to matter.
We agree with that in a few specific situations:
- Your budget is small. In the accounts we manage, a sales campaign spending less than about $50 a day rarely has enough conversions to feed two ad sets, and splitting it leaves both halves too thin to learn. Set up the audience segments, watch the breakdown, and only split if existing customers are taking a large share.
- Your customer list is small. If Meta can only match a few hundred people, the most they can absorb is modest. Check the breakdown before building anything.
- The breakdown already looks fine. If existing customers are taking a small single-digit share of spend on their own, the campaign is not drifting and there is nothing to fix.
So the order of operations is: define your audiences, read the breakdown, and let that number decide. A split is a fix for a problem you can see, not a structure every account needs.
What no account should do is leave the question unasked. Meta removed the setting, but it did not remove the reason the setting existed. Somebody has to decide how much of your budget goes to people who already said yes. If you do not, the campaign will, and it will choose whatever makes its own numbers look best.
If you want a second set of eyes on how your budget splits between new and returning customers, that is something we look at in every Meta account we take on. And if you would rather set it up yourself, everything you need is above.
