If you run Meta ads, the last two weeks of headlines have probably reached you in some form. Meta is changing how its Pixel counts sales. A new setting switches on by itself on November 12. Your reports are about to show more conversions. Some of the coverage reads like a warning, and some reads like good news, and almost all of it leaves out the part that decides whether any of it touches your business.
Here is that part. Meta's own instructions for the new setting list the conditions a website has to meet before it does anything at all. One of them is that the site must require people to log in to an account. A typical small business website does not ask anyone to log in. A store with guest checkout does not require it. A plumber's quote form does not. A dental office's booking page does not.
So for a large share of the businesses reading this, the honest summary is short: the switch will be on, and it will do nothing. For the rest, it is worth fifteen minutes this week, because the change lands fifteen days before Black Friday, and a jump in your Meta numbers that week is going to look like your ads suddenly got better.
This piece walks through what the setting does, the three conditions in Meta's own words, how to tell which side of the line your website sits on, and what to write down before November 12 so you can read your numbers honestly afterward.
Meta Matches the Sale on the Shopper's Own Phone. Your Attribution Rules Stay the Same.
The setting is called Measure conversions on customer devices, and it lives in Meta Events Manager, inside your Pixel's settings.
To understand what it fixes, start with the problem Meta describes. Meta's own example is a lamp. Someone sees your ad for a lamp on Instagram. Two days later they come to your website on their own and buy it. Meta may have a record of the ad view, and it may receive the purchase from your Pixel, and still fail to connect the two. When that happens, the sale happened, but no campaign gets credit for it.
Normally that matching happens on Meta's servers. Your Pixel sends the purchase, Meta tries to tie it to a person who saw or clicked an ad, and if the trail has gone cold, the sale goes unclaimed.
With the new setting on, Meta says the match happens on the customer's device instead, "where both the ad view and the purchase are present." It uses information the customer has already shared with you, such as an email address, plus what Meta calls "temporary, non-identifying device information." According to Meta, it then receives only two facts: a purchase happened, and which campaign gets the credit.
Two things do not change, and both matter.
First, it does not replace anything. Meta says it "works alongside your existing Pixel and Conversions API setup and doesn't replace it." If you already send sales to Meta from your server, keep doing exactly that. If you have never checked whether your sales data is reaching Meta cleanly, that is still the bigger job, and we walked through it in Meta Scores Your Sales Data Out of 10.
Second, your attribution rules stay put. In Meta's words: "Your attribution window and attribution model don't change." The setting changes which sales Meta can see, not how it credits them.
Look closely at the lamp example, though. The buyer never clicked. They saw the ad and came back later. That is a view-through sale, and whether Meta counts those at all depends on your attribution setting. If your ad sets count a sale within one day of someone viewing an ad, a sale like this can count. If they count clicks only, a view-only match like the lamp has nothing to count against. That setting has always been an instruction to Meta rather than a neutral report, which we explain in That Attribution Setting Is Not a Report. It Is an Instruction., and this change makes it matter a little more.
Meta Lists Three Conditions. Miss One and the Setting Does Nothing.
Meta's help page for the feature has a short section titled "Before you turn on measuring conversions on customer devices." Here is what it says, word for word.
- "You must install the Meta Pixel on your website. Install the pixel directly on your page, not inside an iframe or as an image-only (img) pixel."
- "Your website must require people to log in to an account," using an email address or a phone number.
And at the top of the same page: "Measuring conversions on customer devices is not active for people located in the EU and UK."
That gives you three conditions, and a sale has to clear all three.
The Pixel sits directly on the page. If your Pixel was installed through your store platform, a plugin, or a tag manager, it almost certainly runs on the page itself, which is what Meta wants. The cases that fail are narrower: a Pixel that only fires inside an embedded frame (common with booking tools and some form builders that load from another company's website into a box on yours), or an old image-only Pixel. A business that sends its sales to Meta only from a server, with no Pixel in the browser at all, also fails this one.
The site requires a login. This is the condition that rules most small businesses out, and it is the one the headlines skipped.
The buyer is outside the EU and UK. For most of our readers this is a non-issue. It applies per buyer, so a US store with a handful of UK customers gets nothing on those UK sales.
One more thing about the date. November 12 comes from an email Meta sent to advertisers, which trade publications reported in mid-October. Meta's help pages describe the feature but do not give a start date. Plan around November 12, but do not be surprised if the rollout is gradual.
A Contact Form Is Not a Login
Read the login condition again: the website must require people to log in to an account with an email or phone number. Not offer an account. Require one.
Now picture the websites most small businesses actually run.
A roofing company has a quote form. The customer types a name, phone number and address, and nobody creates an account. That is not a login.
A med spa books appointments through a scheduling tool that appears in a box on its website. The customer never logs in to anything, and the booking tool very likely runs in an iframe, which fails the first condition as well.
A home services company runs call ads. The customer taps a button and the phone rings. No website visit at all, so nothing here applies.
A boutique sells online and lets people check out as guests, which most small stores do because every extra step at checkout costs sales. Nobody is required to log in.
On all four of those sites, the new setting will show as on in Events Manager and will measure nothing extra. That is not a problem. It just means the news is not about you, and you should not expect your numbers to move because of it.
The businesses it can actually reach are the ones where an account is part of buying:
- Stores that make customers create an account before they can check out
- Subscription and membership businesses, where customers log in to manage what they pay for
- Gyms, studios and clubs that sell or book through a member login on their own website
- Wholesale or trade-only stores, where every buyer has to be approved and signed in
There is a gray zone, and it is a common one. Plenty of stores offer an account but do not require it, so some buyers sign in and others check out as guests. Meta has not said how it treats that setup. The plain reading of "must require" suggests guest buyers are not covered, and we would not plan around anything better than that until Meta says otherwise.
Checking your own site takes five minutes. Open your website in a private browser window, add something to the cart or start a booking, and see whether you can finish without creating an account or signing in. If you can, the login condition is not met. Then open your Pixel's settings in Events Manager and look at whether the new toggle appears, and what it says.
Meta's Sales Number Goes Up. Your Customer Count Stays Exactly Where It Was.
If your site does qualify, here is what to expect, and here is the trap inside it.
The sales Meta "recovers" are sales that already happened. A customer saw an ad, came back later and bought. Before November 12, that sale showed up in your store and your bank but not in Ads Manager. After November 12, some of those sales start appearing in Ads Manager too, credited to a campaign.
Nothing about your business changed. You have exactly as many customers as you would have had. But every number in Ads Manager that is built on conversions moves at once. Purchases go up. Cost per purchase goes down. Return on ad spend goes up. If you glance at the dashboard in the third week of November, it will look like your ads just improved, right as you are deciding how much to spend on Black Friday.
That is the moment to be careful. Raising a budget because a counting change made the ads look cheaper is how owners overspend in the most expensive weeks of the year. Meta already tends to report more sales than your own records show, for reasons we unpack in Meta Says It Made You 40 Sales. Your Bank Says 12., and this change can widen that gap without a single extra order behind it.
Meta also says the feature can "improve campaign performance with a stronger signal," which would mean the extra matches help its system find buyers, not just report them. It has not said whether these recovered sales feed bidding and delivery, or whether the sale amount comes through with them. Until it does, treat any movement in your reported results after November 12 as partly a change in counting, and judge your ads by your own orders.
Logged-In Matching Is Fine for a Shoe Store. If You Sell Anything Medical, Turn It Off.
So should you leave it on? For most businesses that qualify, yes.
The case for leaving it on is simple. It replaces nothing, it does not change your attribution rules, and if it does feed Meta's delivery system, better information about who actually buys is the thing that system runs on. Yiqi Wu, who runs a company that builds Pixel tracking for Shopify stores and so has a stake in this, put it plainly in a post on X: "I wouldn't opt out by default." His one reason to switch it off was a store whose Meta numbers stop matching its real orders with no explanation.
We agree, with two exceptions.
Health and other sensitive businesses. If you run a medical practice, a therapy practice, a med spa with medical services, a pharmacy, or anything in finance, the question is not only about reporting. Meta's help page says it is your responsibility to make sure your setup does not share information Meta prohibits, and Meta has not published how this feature handles a visitor who declined tracking. A login on a health website usually sits in front of exactly the pages where sensitive information lives. MetricMaven, which publishes measurement guidance, recommends that banks and healthcare organizations switch the feature off until their legal and privacy advisers have reviewed it. That is the position we would take too. If you are in this group, you already know Meta treats you differently, and we covered what that means in Your Health Business Cannot Retarget People the Normal Way.
Anyone who cannot check the count. If you have no way to compare Meta's reported sales with your actual orders, you cannot tell a counting change from a real one. Fix that first. Switching the setting off will not make your numbers right; it only removes one source of movement while you set up the comparison.
Turning it off, or confirming it is on, takes a minute:
- Open Meta Events Manager.
- Choose Data sources from the menu.
- Select your Pixel.
- Click Settings.
- Under Meta Pixel settings, find Measure conversions on customer devices and set it the way you have decided.
Meta says you can turn it off at any time. Whichever way you set it, write the decision and the date somewhere your future self will find it.
After November 12, the Only Clean Baseline Is the One You Saved This Week.
Whether or not your site qualifies, the most useful thing you can do before November 12 is write down where your numbers are right now. Once the change rolls out and Black Friday arrives on top of it, you will not be able to reconstruct a clean "before" from the dashboards. Ads Manager will happily show you old dates, but it will not tell you what changed underneath them.
Here is the fifteen-minute version.
- Export the last 30 days from Ads Manager. Daily purchases (or leads, or whatever you optimize for) by campaign. Write down your attribution setting next to the export, because the numbers mean nothing without it.
- Pull your real orders for the same 30 days. From your store, your booking system, your CRM, or your bank. Count only orders that came in online, since that is what the Pixel can see.
- Divide Meta's number by yours. If Meta says 45 sales and your store shows 60 online orders, your ratio is 0.75. If Meta says 70, it is 1.17. There is no perfect number. What matters is that you know yours.
- Mark November 12 on every report you look at. A sticky note on the monitor counts. The point is that nobody in your business reads a jump in late November without remembering why it might be there.
- Change nothing else in your tracking that week. If you add a new plugin or switch on a new data setting at the same time, you will never know which change moved the numbers.
After November 12, keep calculating the same ratio each week. If Meta's number rises and your orders do not, the difference is counting, not customers.
The chart above is an illustration, not a client account, but it shows the shape to watch for: a ratio that sits steady for weeks and then steps up on a date with nothing new happening in your store.
If you want to know what your Meta ads actually cause, as opposed to what they get credited with, the answer is a holdout test, where Meta withholds ads from part of the audience and compares the two groups. That is a bigger job, and it needs more sales than many small businesses have, which we cover in Meta's Real Sales Test Needs 500 Conversions. Don't start one this month. Wait until after the holidays, when the counting change has settled and the season is not distorting everything at once.
Most Sites Can Ignore November 12. Login Sites Need One Saved Number First.
Most of what has been written about this change assumes every advertiser is about to see more conversions. Meta's own conditions say otherwise. If nobody has to log in to your website, you can confirm that in five minutes and stop thinking about it. If customers do log in, leave the setting on unless you are in health or another sensitive category, save your baseline this week, and judge November by your orders rather than by Ads Manager.
Checking settings like this one, and keeping the reported numbers honest against the real ones, is a big part of what we do in Meta ads management. If you would rather hand that off and focus on the busiest month of your year, we are happy to help. If you would rather do it yourself, everything you need is above.
