Meta reports a conversion. Your CRM shows nothing. This mismatch is not a glitch — it's the normal state of Facebook ads attribution in 2026, and most startups are making budget decisions on numbers they don't fully understand.
Why Facebook Ads Attribution Has Become More Complicated Since iOS 14
iOS 14 broke the Meta Pixel's ability to track user behavior across apps. Meta's response: introduce Conversions API (CAPI) for server-side tracking, and model the conversions they could no longer directly observe. The issue is that modeled conversions are presented in Ads Manager with the same visual weight as directly measured ones — no asterisk, no confidence interval.
Consent Mode v2 affects what data you can measure; see our Consent Mode v2 guide.
The result: Meta-reported conversions routinely overcount. For some accounts, Meta reports 3-4x the conversions that appear in the CRM.
7-Day Click vs 1-Day Click: Which Attribution Window to Use
Meta's default is 7-day click, 1-day view. View-through attribution is the main driver of inflation: someone sees your ad, ignores it, searches for your product, and converts via organic — Meta claims credit.
For most B2B SaaS startups with sales cycles longer than a week, 1-day click is the most defensible window. It reduces channel overlap and correlates more closely with CRM data. The tradeoff: campaign performance looks worse in Ads Manager, which is exactly why most account managers default to the broader window.
How Meta Pixel, CAPI, and Modeled Conversions Work Together
Meta Pixel fires browser-side. Its coverage dropped post-iOS 14 but remains useful for non-iOS browsers and consenting users.
Conversions API (CAPI) fires server-side directly from your backend or CRM, bypassing ad blockers and iOS restrictions. CAPI is now the primary signal source for accurate attribution.
Modeled conversions fill gaps where neither pixel nor CAPI can observe behavior. These estimates appear in Ads Manager totals without being labeled as estimates.
The Gap Between Meta-Reported Results and Actual Business Outcomes
A simple calibration process: (1) pull Meta-reported conversions for a 30-day period; (2) pull CRM data filtered by paid social touchpoints; (3) calculate the ratio. For most B2B SaaS accounts, expect Meta to report 1.5-3x the CRM number. That ratio becomes your attribution discount factor. Apply it when evaluating campaign profitability.
How Agencies Set Up Attribution for a Trustworthy View
CAPI with deduplication. Server-side events sent via CAPI must include event IDs that match the pixel's browser-side events. Without deduplication, you count the same conversion twice.
UTM parameters on every ad. All Meta traffic should carry UTMs that land in your analytics platform, creating a parallel measurement stream Meta cannot influence.
Monthly CRM reconciliation. Compare Meta-reported pipeline against CRM-sourced pipeline for the same period. Track the ratio over time.
Frequently Asked Questions
What Is the Best Attribution Window for Facebook Ads?
For most B2B SaaS startups, a 1-day click window provides the most accurate view. The default 7-day click + 1-day view window inflates results by including view-through conversions that may have converted through other channels.
Why Do Facebook Ads Show More Conversions Than My CRM?
Meta combines directly tracked events with statistically modeled conversions to fill gaps from iOS tracking restrictions. View-through attribution also credits conversions where the user only saw the ad. The combined effect regularly causes 1.5-3x over-reporting vs. CRM data.
What Is Conversions API and Do I Need It?
CAPI sends conversion data directly from your backend to Meta, bypassing browser-level restrictions. Without it, a large share of your conversions rely entirely on Meta's statistical models, which reduces reliability.
Key Takeaways
- iOS 14 forced Meta to rely partly on modeled conversions, which inflate Ads Manager totals without being labeled as estimates
- The 7-day click + 1-day view default is the largest driver of attribution inflation — switch to 1-day click for more accurate reporting
- CAPI is now required for reliable attribution; without it, numbers are heavily model-dependent
- Always verify Meta results against CRM data — build a monthly reconciliation habit
- Apply an attribution discount factor when comparing Meta's reported CPA to actual acquisition cost
- Platform-reported vs. reality gaps are widest on Meta and must be accounted for in any channel comparison
Building a Unified Attribution View Across Channels
A single platform report never tells the whole story. When Meta, Google, and LinkedIn each model their own conversions, the only honest number is the one you build yourself from first-party CRM data. Export closed-won revenue by original touchpoint, then weight each channel by its assisted-conversion share rather than its last-click credit. This shift from platform-reported attribution to revenue-attributed modeling is what separates startups that scale profitably from those that chase vanity ROAS.
Practical setup: store a persistent click_id on every lead, stamp it at form fill, and reconcile it against the CRM weekly. Over a quarter you will see which channels consistently assist deals even when they never get last-click credit. Reallocate budget toward assisted channels before they starve. Attribution is not a one-time audit; it is a standing operating cadence.
Surviving Signal Loss with Server-Side Data
iOS 14.5 and later throttled the pixel, but the advertisers who recovered fastest were the ones that moved event collection server-side. When a purchase, signup, or trial-start fires from your backend and is sent to Meta through the Conversions API with a stable event ID, you stop depending on a browser that may drop the ping. The practical checklist: send the same event ID from both pixel and CAPI so Meta deduplicates instead of double-counting, pass hashed email or phone where you have it to improve match rates, and test the events in Meta's Events Manager before scaling spend behind them.
Server-side tracking also protects you from ad blockers and browser privacy changes that will keep arriving. Treat CAPI as the system of record and the pixel as a convenience layer. Once that foundation exists, the modeled-conversion gap shrinks because Meta has more real signal to model from, and your CRM reconciliation ratio moves closer to 1:1.
When to Trust Meta and When to Trust Your CRM
Use Meta's numbers for optimization - bidding, audience, and creative decisions - because Meta has the freshest in-platform signal. Use your CRM for truth - profitability, payback, and pipeline. The mistake is letting one stand in for the other. A campaign that looks unprofitable in Ads Manager may be your best pipeline source once the attribution discount is removed, and a campaign that looks great in-platform may be stealing credit from organic. Hold both views, reconcile monthly, and let the reconciliation ratio, not the raw Meta total, drive budget calls.
A Simple Weekly Attribution Routine
Make reconciliation a habit, not a project. Every Monday, pull Meta-reported conversions for the prior week and the matching CRM-closed leads by touchpoint, then compute the ratio. Plot it on a running chart so you can see drift before it becomes a budget problem. When the ratio climbs above your historical norm, tighten attribution windows and revisit CAPI coverage; when it falls, you have earned room to scale spend. Fifteen minutes a week keeps the whole account honest and prevents the slow over-counting that quietly drains a budget.
Documenting Your Attribution Method for the Team
Write the discount factor and reconciliation process into a one-page internal note so the next analyst does not rediscover it. State the attribution window you use, the CAPI setup, and the current Meta-to-CRM ratio. When someone questions why a campaign looks unprofitable in Ads Manager yet keeps its budget, the note explains the reasoning in fifteen seconds instead of a meeting. Attribution only protects the budget when the method is shared and stable, not locked in one person's head.