Illustration: two sample bars labelled reported in ads and orders in the store, with the gap between them marked, and a row of checks with one flagged in red, with a sample data label

Insight · November 17, 2026

Meta now counts more sales. Are they real sales?

Meta’s on-device purchase counting is on by default, and its settings panel shows a November 12 start. Here is what Meta says, what it has not said, and how to keep your numbers honest before Black Friday.

  • Setting on by default
  • Panel shows Nov 12 start
  • Not active in the EU and UK
  • Pixel and CAPI stay
  • Attribution window unchanged
  • Compare with real orders

The story

Reported sales are about to change. Real ones are not

  • November 17, 2026
  • Meta Business Help Center, PPC Land (October 27, 2026) and Stape (October 22, 2026)
  • Insight
  • By Silas Node, AONE Creative

A setting in Meta Events Manager called “Measure conversions on customer devices” is meant to measure conversions that Meta’s reports would otherwise miss, and its settings panel shows measurement starting November 12, 2026, as PPC Land and Stape report. Meta’s Business Help Center says the setting “is turned on by default,” that it is “not active for people located in the EU and UK,” and that the system measures conversions on the customer’s device, using information they already shared with the business (like an email) and temporary, non-identifying device information. Meta’s example: someone sees a lamp ad on Instagram, visits the website two days later and buys it. Meta may have a record of the ad view and the purchase event but cannot always connect them. With the setting on, that match happens on the device, and “Meta only receives confirmation that a purchase happened and which campaign to attribute it to.” Meta’s help page itself gives no date. PPC Land reports that advertisers can switch the setting off at any time in each dataset’s settings.

Meta says the setting works alongside the existing Pixel and Conversions API setup and does not replace it, and that “your attribution window and attribution model don’t change.” It “affects which conversions can be measured, not how they’re credited to your ads.” Meta’s page says keeping the setting on will “improve campaign performance with a stronger signal,” and does not say how. PPC Land reports that Meta has not said whether recovered conversions feed bidding, whether purchase values are included, or how consent is handled, and adds that MetricMaven lists as unresolved whether they feed delivery, audiences or only reporting. Stape warns that reported cost per acquisition may fall and return on ad spend may rise even if real sales and spend stay the same, because more existing purchases get credited to ads. More attributed conversions does not necessarily mean more actual conversions. Meta’s page also puts a duty on the advertiser: “It’s your responsibility to make sure your integration doesn’t share prohibited information with Meta.” We have no connection to Meta, and this account rests on the pages named above.

An example of how it could work. Sample Ridge Outfitters is a fictional online store that sells camping gear and spends a steady budget on Meta ads. For the four weeks before November 12 it writes down two numbers every day: purchases Meta reports, and paid orders in its own store. The two numbers track closely. After the change, a sample week shows 130 reported purchases against 100 orders. Nothing about the ads changed, and the store did not sell more. The team does not cut the budget in a panic or raise it in delight. It keeps its own order count as the number it trusts, compares the two every day, and asks Meta’s reports to explain the gap before anyone moves money.

Theories, in our view. First, a jump in reported results after November 12 may be a change in counting, not in demand, so the store’s own orders are the check that matters. Second, the timing is awkward: the changed baseline arrives about two weeks before Black Friday on November 27, when budgets are largest and mistakes cost the most, and a buyer who judges this month against last month is comparing two different rulers. Third, it may be wiser to leave the setting on and read it with care than to switch it off by default, though that depends on the account. If the numbers stop matching orders and no one can find why, switching it off is a fair test. These are our views, not findings from the sources above.

How AONE Creative would set it up. Our media buying covers paid campaigns on the major social networks alongside digital, cable TV and radio, and it includes reporting on what ran, where and when. For a change like this we would write down the platform numbers and the store’s own orders before the switch, keep a daily comparison of the two, and mark the date of the change in every report so nobody reads a counting change as a result. Meta’s page claims a performance benefit but, as reported, leaves open how recovered conversions are used, so we would not promise what they do for performance. We would measure what they do for your numbers.

Where to start. Book a strategy call and bring your last eight weeks of platform reports and your own order counts. We will look at them together and agree on what to compare and when. If you also want AI to find what else your team repeats by hand, the AI Readiness Audit is Snapshot $1,500 or Blueprint $4,500.

Share this story

Join the conversation

If Meta reported 30% more sales tomorrow, would you know how many were real?

Send us your answer, or book a call and we will compare your numbers together.

What this means for you

Compare the two numbers before the sale

Bring your platform reports and your order counts, and we will tell you what we would compare.

Related

  • Service

    Media Buying

    Ad buys and delivery across digital, social, cable TV and radio, with reporting on what ran.

  • Digital campaigns built and measured as one push.

  • Service

    Web Platforms

    Sites and stores built to send clean purchase data.

  • A map of how your team works and the automations worth building. Snapshot $1,500 or Blueprint $4,500.

Start with AI

Put AI to work on one process

An AI Readiness Audit maps how your team works and ranks the automations worth building, whether or not you hire us to build them. Snapshot $1,500 or Blueprint $4,500.

A camera crane reaches across the Paint Misbehavin' burn set with a fire truck standing by, 2009, in black and white

Insight

Which number do you trust on Black Friday?

Tell us where you buy and how you count orders. We will tell you where we would start.