One of the most common questions when working with Facebook Conversion API (CAPI):

— in adset.pro conversion refers to a specific ad set/creative
— and in Facebook Ads Manager it suddenly appears under another one 😅

At first glance, it seems like a bug or that the attribution has “broken.” In practice, it’s simpler: the tracker and Facebook answer different questions — hence the numbers may differ.

Two Logics, Two Truths

1. How the Tracker Calculates (adset.pro)

The tracker lives by clicks.
We record the original click (Click ID), save the event chain, and tie the conversion to where the user actually came from — to a specific transition/source/creative/ad set (via macros, fbclid, UTM, your Click ID, etc.).

In adset.pro conversion = the result of a specific click.

2. How Facebook Calculates

Facebook lives by attribution and delivery models.
For it, the last touches of the user within the attribution window are important — and these can be:

• click (click-through)
• impression (view-through) — the user simply saw the ad in the feed, did not click

In Ads Manager, conversion = the result of the “last interaction” in Meta’s logic.

---

What This Means in Practice

Let’s imagine a real scenario:

1. The user clicks on ad A → lands on your track-URL → then the funnel (landing / PWA / application / installation / purchase).

2. After a while, they open Facebook again “just to scroll through the feed.”

3. The algorithm shows them ad B (another creative/ad set). The user may not click anything, just see it.

4. The conversion happens later (immediately or after some time).

In adset.pro the conversion will remain tied to the click on ad A — because the tracker knows: “the entry was from here.”
And Facebook may attribute the conversion to ad B, because for it, the last interaction (even an impression) became more “weighty” in the attribution model.

This is where the “discrepancy” arises: the same conversion is distributed according to different rules.

---

Practical Conclusion for Media Buying

If a user clicked on one ad, and another just saw it in the feed, then for management decisions, the priority is clear:

✅ rely on the analytics of adset.pro, because this ties to a specific click and your funnel
⚠️ and not on how Facebook “redistributed” the event within its attribution model.

That’s why sometimes you see the situation:

• in Ads Manager “the ad set is empty”
• while in the tracker the ad set actually generated clicks → leads → purchases

---

Why This Is Dangerous for Auto-Rules

If you use auto-rules based solely on Ads Manager, you might accidentally:

• turn off an ad set that actually brought conversions per click
• keep an ad set that “picked up” view-through, but was not the source of entry

In other words, auto-rules without a tracker can cut off working connections.

---

What You Can Do Right Now

1. Mark “Risk Zones” in Naming

The simplest way is to mark ad sets where view-through/reallocations often occur:

For example:
FR_M_A1+ or FR_M_A1_vt+

This is trivial, but helps the team avoid making hasty decisions when Facebook statistics “go awry.”

2. Separate Decision-Making:

• Optimization of creatives/connections/sources → by adset.pro (click logic)
• Checking delivery/signals within Meta → by Ads Manager (Meta attribution)

3. Carefully Interpret “Conversions Attributed by Facebook Incorrectly”

Should you turn off the ad set to which Facebook attributed a conversion “by last contact”?
There is no definitive answer.

Skeptical approach:
“There’s little benefit from it — the user would have converted anyway after the first touch.”

Pragmatic approach:
“If Meta shows this user exactly this ad — then the algorithm finds people likely to convert. Perhaps the ad set helps catch up and strengthen.”

In such cases, we recommend:

• not to cut off immediately
• give it some time/limited budget
• look at the connection as a whole: entries/quality/follow-up/repeated touches

---

Main Point

The discrepancy is not always an error. Most often, it’s a difference in philosophies:

• tracker = “where the user came from”
• Facebook = “how Meta attributed the event in its model”

And if your task is to manage real ROI, then the basis for decisions should be the click analytics of the tracker (adset.pro), while Facebook reports are an additional layer that is useful but does not always explain the cause-and-effect relationship.