Meta AdsModule 4: Measurement, automation and responsibilityLesson 10 of 11
Course progress82%
20 min lesson · Updated August 2026
How does Meta attribution work?
Meta reports delivery, response and conversions attributed within selected windows; these figures are modeled platform evidence, not a complete or causal account of every sale.
What you will learn
By the end, you will understand:
Calculate reach, frequency, CPM, CTR, CPC, CPA and ROAS
Explain click/view attribution windows
Reconcile Meta reporting with analytics, CRM and orders
Visual explainer
See the idea clearly.
♥
A campaign reaches 8,000 people with 12,000 impressions, receives clicks and view-through exposure, then attributes selected purchases while the store records all orders separately.
Delivery metrics describe exposure
Metric
Meaning
Formula
Reach
Estimated number of people/accounts exposed under Meta’s method.
Impressions
Times ads were shown; one person can receive several.
Frequency
Average exposures per reached person.
Impressions ÷ reach
CPM
Average cost per thousand impressions.
Cost ÷ impressions × 1,000
Response metrics need precise definitions
CTR is clicks divided by impressions, but Meta can report link CTR, outbound CTR and all-click CTR. Use the column that matches the question. CPC likewise depends on the click definition.
A landing-page view requires the destination to load and is different from a link click. Neither proves that the visitor became a customer.
Outcome formulas
Metric
Formula
Limit
Cost per result / CPA
Cost ÷ attributed results
Depends on the selected result and tracking.
Conversion rate
Conversions ÷ relevant interactions × 100
State the denominator and attribution source.
ROAS
Attributed conversion value ÷ ad cost
Not the same as profit.
Attribution windows define eligible credit
A person may click or view an ad and convert later. The attribution setting defines which interactions and time windows can receive reporting credit. Available click-through and view-through windows can vary by objective and account.
A view-through conversion can be useful evidence of exposure contribution, but it has more causal uncertainty than a direct click path. Use breakdowns where available and document the window.
Click-through
A person clicked an eligible ad and converted inside the selected window.
View-through
A person was shown an eligible ad without clicking, then converted inside the selected window.
Meta and analytics see different journeys
Meta may use account matching, modeled conversions and view-through attribution. Web analytics may require a session, consent and identifiable campaign click. Stores or CRMs record completed outcomes but may not know which ad influenced them.
Differences are expected. Align time zones, currencies, event definitions and date-of-click versus date-of-conversion reporting before investigating gaps.
Attribution is not incrementality
An attributed purchase means it met the platform’s attribution rules. It does not prove the purchase would not have occurred without the ad. Existing customers and brand demand can create credited sales that are not entirely incremental.
Use suitable lift tests, holdouts, geographic experiments or marketing-mix evidence where scale and decisions justify causal analysis.
A trustworthy report
Names the objective, result and attribution setting.
Separates reach, interaction and business outcome metrics.
Checks qualified leads, refunds, margin and repeat customers.
Reconciles Meta, analytics and business systems directionally.
Labels modeled or estimated data.
Avoids claiming causation from attribution alone.
Real-world example
Example: why three systems disagree
Example
Meta reports 120 purchases using click and view attribution. Web analytics reports 83 paid-social session purchases after consent and detectable clicks. The store records 150 total purchases from all sources, including refunds later. None is automatically wrong; each answers a different question. The team aligns definitions before deciding budget.
Try this
Annotate one report
Beside every headline metric write its formula, result definition, attribution window, data source and whether it is observed, modeled or downstream verified. Remove any claim that the evidence cannot support.
Common questions
Questions beginners ask.
What is frequency?
Impressions divided by reach: the average number of exposures per reached person.
What is CPM?
Cost per thousand impressions.
Which CTR should I use?
Use link or outbound CTR when measuring destination response, and state the chosen definition. All-click CTR includes other interactions.
What is view-through attribution?
Credit for a conversion after an ad impression without an eligible click, within the selected window.
Why is Meta ROAS higher than analytics?
Meta may include view-through, modeled and cross-device/account-matched conversions under different attribution rules.
Is high frequency bad?
Not automatically. Interpret it with audience size, objective, creative fatigue, response and reach needs.
Does attributed ROAS equal profit?
No. It normally excludes costs and may include outcomes that are not incremental.
How can incrementality be measured?
Through suitable randomized lift/holdout tests or other causal methods when scale and design allow.
