twimco.Let’s talk ↗︎
Menu
Field notes / Commerce

Meta may start counting more of your ad sales tomorrow. Write down today's numbers first

From November 12, your Meta reports may show more purchases with no extra orders behind them. A new Pixel setting is on by default for eligible pixels. Here's who it reaches, what it does and doesn't change, and the ten minutes worth spending today.

Editorial engraving of a bright loft office above an apparel brand's stockroom in daylight: a gray-haired woman in a sage cardigan holds a printed order report and presses a coral flag onto one square of a blank wall planner, while a younger colleague at a laptop with a calculator and an open ledger works through the numbers beside her.

Meta's new Pixel setting, "Measure conversions on customer devices," is on by default for eligible pixels. Events Manager and Meta's mid-October advertiser notice, as reported by PPC Land and Wortfilter, put the start at November 12, 2026, with Wortfilter describing a gradual rollout. It can add purchases Meta was missing to your reports without any new orders. Leave it on for most stores, but export Meta purchases next to real orders by day today, mark the date in every report, and switch it off per pixel if you're mid-test, in a regulated sector, or the numbers stop matching.

Thursday's report, before you open it

Here is a small scene that will play out in a lot of stores this week. On Thursday morning someone opens Ads Manager and the cost per purchase looks better than it did on Wednesday. Nobody changed a campaign. The shop didn't sell more. Somebody still writes "Meta is working" in the weekly update.

The reason is a new Meta Pixel setting called "Measure conversions on customer devices." Trade write-ups that quote Meta's help pages say it's on by default for eligible pixels, so if you haven't touched it, it's probably on. PPC Land reported that under the toggle in Events Manager, Meta shows one grey line: "Measurement starting November 12, 2026." That's tomorrow. The same article reported Meta's mid-October notice to advertisers, which first circulated from October 18. Wortfilter, describing that email, said Meta would roll the feature out gradually from that date.

In plain words, Meta sometimes knows a person saw your ad and later receives your order event, but can't connect the two. Meta's help article, as quoted below, gives the example of someone who sees a lamp on Instagram and buys it on your site two days later. With the setting on, Meta tries to make that match on the shopper's own phone or computer, using information they already gave you (like an email address) and temporary device data. Meta says it learns only that a purchase happened and which campaign gets the credit.

So your reports may show more purchases from tomorrow. Some of those will be purchases Meta wasn't crediting to an ad. None of them are new orders. That's the whole point of this note: write down where your numbers stand today, so you can tell those two things apart later.

Below are the questions we'd expect a store owner to ask, in the order they usually come up.

Decision guide

What changes on November 12

ConsiderationBefore November 12From November 12, setting on
Where the match happensOn Meta's serversOn Meta's servers and on the shopper's device
Purchases in your reportsThose Meta can connect on its serversThose, plus ones matched on the device
Attribution window and modelYour current settingsUnchanged, per Meta
Pixel and Conversions APIRunningStill running; the setting doesn't replace them
Your real ordersWhat your store soldExactly the same
Summary of Meta's help pages as quoted by Adsap and Relevant Audience, and of the Events Manager panel as reported by PPC Land; reading aid, not Meta documentation.

Does this even apply to my store?

Maybe not, and it's worth knowing which. Meta's help pages, as quoted by Adsap and Relevant Audience, set three conditions.

First, where the shopper is. The feature isn't active for people located in the EU and UK. That follows the shopper, not your business, so a European brand still qualifies for its American visitors.

Second, how the Pixel is installed. It has to sit directly on your pages, not inside an iframe and not as an image-only pixel.

Third, how people use your site. Meta's wording is that your website must require people to log in to an account with an email or mobile phone number. Read literally, a store where most people check out as guests may see little change. Meta doesn't say how it checks this, or whether an email typed at checkout counts. So don't assume either way. Look at the numbers after the date.

If customers sign in to buy from you, through store accounts or a loyalty program, assume it may apply. Either way, open Events Manager, go to Data sources, pick your pixel, open Settings, and see whether the toggle is there and what the grey line says. That takes two minutes, and it's more reliable than any article, this one included.

The stores that get burned by this aren't the ones that leave the setting on. They're the ones that read Thursday's report without knowing anything changed.

So are my ads suddenly working better?

Not because of this. Counted and caused are different things.

A conversion recovered by this setting is a purchase that already happened, which Meta now ties to an ad. Those same help-page write-ups say Meta states that your attribution window and attribution model don't change. What changes is how many purchases it can see. Whether the ad actually caused that purchase is the same open question it always was.

There are also a few things Meta hasn't said, and they matter for how your reports will look:

  • whether the order value travels with a recovered purchase, which decides whether return on ad spend moves or only the purchase count,
  • whether recovered purchases feed bidding and delivery, or only reporting (Meta's case for leaving it on includes a "stronger signal," which suggests they do, but it doesn't say how),
  • how a device match is reconciled with a Conversions API event for the same order.

None of that is a reason to panic. It is a reason not to compare last week's cost per purchase with next week's as if nothing happened in between. Black Friday is November 27, fifteen days after that date. Holiday budgets get set on exactly those comparisons.

Should I switch it off?

For most stores, no. Leave it on, write down your baseline first, and keep an eye on it. If Meta has been missing real sales, you want them counted.

There are three situations where switching it off for now is the careful choice.

You're in the middle of a test. If you started a creative or budget test before November 12, it needs one way of counting until it ends. Changing the measurement halfway through makes the result hard to trust.

Your business has privacy rules of its own. Health, finance, anything with sensitive customer data. A new measurement path running on customer devices is something whoever owns privacy should hear about before it runs, not after. The help pages, as quoted, don't mention consent banners.

Your numbers stop matching and nobody can explain why. If Meta's purchase count climbs and your real orders don't, and you can't find the reason, turn it off and look again.

The switch is per pixel: Events Manager, Data sources, your pixel, Settings, then "Measure conversions on customer devices." If you have more than one pixel, do each one. If an agency runs your ads, ask them one question today: which pixels is it on for, and how will next week's report show the change? "We've handled it" isn't an answer.

This is the second default Meta has flipped on advertisers in a month. On October 28 it turned on product browsing in running ads unless people opted out. The habit is the same both times: decide on purpose, and keep a record of what changed and when.

How the work moves

Ten minutes before November 12

  1. Check the toggle

    Events Manager, Data sources, your pixel, Settings. Is it there, and is it on?

  2. Export a baseline

    Meta purchases and paid orders by day for the last 30 days, with the daily ratio.

  3. Mark the date

    Note November 12 on every report that shows Meta cost per purchase.

  4. Decide on purpose

    Leave it on, or switch it off per pixel if you're mid-test, regulated, or can't reconcile.

  5. Watch to Black Friday

    Recheck the ratio every few days through November 27.

Illustrative checklist; not a client example or measured result.

Ten minutes before tomorrow

You don't need a tracking project for this. You need a baseline and a note on the calendar.

  1. Export two columns. Meta-attributed purchases by day, and your store's paid orders by day, for the last thirty days. Put them side by side in a sheet and work out the ratio for each day.
  2. Mark November 12 in every report. Any weekly update, dashboard or agency deck that shows Meta cost per purchase should carry a note that counting may change from that date.
  3. Keep the Conversions API running. Those help-page write-ups say the new setting works alongside the Pixel and the Conversions API and doesn't replace them. If your server-side events are healthy, leave them that way.
  4. Check the ratio every few days through Black Friday. If Meta's count rises and your orders stay flat, the measurement changed and sales didn't. If both rise, great, but that's when a holdout test or a look at new customers earns its keep.

That's it. The stores that will get burned by this aren't the ones that leave the setting on. They're the ones that read Thursday's report without knowing anything changed.

If your store's tracking, checkout or reporting needs a hand before the season, that's work we do. See our e-commerce work or get in touch.

Related work: Paddle8.

twimco.

We build alongside the people doing the work.
Bring us the problem you’re thinking about.

Related work

Paddle8 ↗︎