How to Reconcile Platform-Reported Performance With Closed Revenue

What this guide helps you do

Ad platforms are useful, but they do not close your books.

Meta, Google, TikTok, LinkedIn, your CRM, Shopify, Stripe, and finance can all report different answers to the same question: what did we acquire last month?

That does not automatically mean anyone is lying. Each system is counting from its own position. The problem starts when the business uses a platform number as if it were closed revenue.

This guide gives you a monthly reconciliation workflow. The goal is to compare what platforms reported against what the business actually closed, identify the gap, and decide what is safe to change.

Use it before budget increases, agency reviews, board updates, channel cuts, and AI-generated performance summaries.

Define the system of record first

Before pulling platform data, decide which system gets to answer the question “what closed?”

Examples:

Business type	Likely system of record
Ecommerce	Shopify, WooCommerce, ERP, finance export
SaaS	Stripe, subscription system, CRM closed-won
Lead generation	CRM closed-won, qualified opportunity, booked appointment plus close status
Services	CRM closed-won, signed agreement, invoice, finance system
Marketplace	Order database, payout system, finance ledger

Write that source at the top of the reconciliation file. If the team cannot agree on it, stop there. That is the first operating problem.

The monthly reconciliation table

Create one worksheet with these tabs:

1. system_of_record
2. platform_exports
3. mapping_rules
4. reconciliation
5. decision_log

Keep it boring. A spreadsheet is fine. The point is repeatability, not software elegance.

Step 1: Pull closed revenue or closed customers

From the system of record, export the period you are reviewing.

Minimum fields:

Field	Why it matters
customer, order, or deal ID	Prevents duplicate counting
close/order date	Aligns the time period
revenue	Shows the business result
margin if available	Prevents scaling unprofitable volume
refund/cancel status	Keeps bad revenue out of the win count
source/medium/campaign if available	Helps match back to marketing activity
first-touch or last-touch fields if available	Useful, but not always complete

Clean the export before comparing it. Remove test orders, duplicates, canceled deals, refunded orders if the decision depends on net revenue, and records outside the period.

Do not average your way through bad records. Flag them.

Step 2: Pull platform-reported performance

Export the same period from each platform.

Minimum fields:

Field	Why it matters
platform	Separates source claims
campaign	Lets you inspect budget decisions
spend	Needed for CAC or ROAS
conversions	Platform claim
conversion value	Platform revenue claim
attribution setting/window	Explains part of the gap
conversion action name	Prevents mixing leads, purchases, calls, and signups

Do not combine platforms yet. Keep each platform’s claim separate until the comparison is done.

Step 3: Normalize the definitions

Most reconciliation problems hide in definitions.

Before comparing totals, answer:

• Is the platform counting leads, purchases, calls, form fills, trials, or qualified opportunities?
• Is the business counting gross revenue, net revenue, booked revenue, collected cash, or margin?
• Are returning customers included?
• Are refunds or cancellations removed?
• Are branded and nonbrand campaigns mixed together?
• Is the platform using click-through, view-through, data-driven, or another attribution model?
• Does the sale happen inside or outside the platform attribution window?

Create a mapping_rules tab like this:

Platform metric	Business metric it maps to	Safe for budget decisions?	Notes
Meta Purchase value	Gross ecommerce revenue	Maybe	Needs refunds removed
Google Ads lead conversion	Form submit	No	Not all leads qualify
CRM closed-won revenue	Closed revenue	Yes	Source of truth

This table protects the team from pretending unlike numbers are the same.

Step 4: Compare totals and calculate the gap

Now build the reconciliation tab.

Source	Reported customers	Reported revenue	Spend	CAC	ROAS
System of record	38	$63,500	—	—	—
Meta	24	$41,200	$18,000	$750	2.29
Google	31	$57,800	$21,000	$677	2.75
Combined platform claims	55	$99,000	$39,000	$709	2.54

Then add the gap:

Comparison	Customer gap	Revenue gap	What it may mean
Platform claims vs system of record	+17	+$35,500	Possible double counting, view-through inflation, duplicate conversion actions, or gross-vs-net mismatch

Sometimes platforms overcount. Sometimes they undercount. Both matter.

If the platform undercounts, the team may cut channels that are helping. If it overcounts, the team may scale campaigns that are not producing real business value.

Step 5: Match records where possible

If you have IDs, emails, GCLIDs, UTMs, order IDs, or CRM campaign fields, match records directly.

Use three buckets:

Bucket	Meaning	Action
Matched	Platform and system-of-record both see the conversion	Usually safe to include
Platform-only	Platform reports it, system of record does not show a closed result	Inspect duplicate, lead quality, attribution window, cancellation, or wrong conversion action
Business-only	Business closed it, platform does not claim it	Inspect direct/organic return, long sales cycle, missing tracking, offline conversion upload, or assisted path

This is where the useful work happens. The operator is not looking for a perfect match rate. The operator is looking for a pattern that changes the decision.

Step 6: Classify the reconciliation gap

Use these classes:

Gap class	Description	Owner
Counting gap	Duplicate events, multiple conversion actions, gross vs net mismatch	Marketing ops / analytics
Timing gap	Sale happens before or after the platform window	Growth owner / analytics
Qualification gap	Platform counts a lead the business would not count as qualified	Sales / lifecycle
Revenue gap	Order value, refunds, discounts, or margin are treated differently	Finance / ecommerce
Source gap	UTM, campaign, or source field is missing or overwritten	Marketing ops / CRM owner
Assisted-path gap	Platform influenced demand but did not get direct credit	Growth owner

Assign one owner per class. A reconciliation gap without an owner becomes a recurring meeting topic instead of a fix.

Step 7: Decide what is safe to change

Use the decision log.

Decision	Evidence	Safe?	Owner	Follow-up
Increase Meta retargeting budget	Platform ROAS strong, but system record shows heavy returning-customer overlap	Not yet	Growth owner	Split new vs returning customers
Cut Google nonbrand	Platform underreports closed-won customers with long sales cycle	No	Sales ops	Upload offline conversions / inspect assisted path
Reduce campaign with high lead volume	CRM shows poor qualification and low close rate	Yes	Paid media lead	Lower bid or tighten targeting

This is the difference between reporting and operating. The table has to protect decisions that are not ready.

Step 8: Use AI to accelerate the review

AI can help once the tables exist.

Good prompts:

textReview this reconciliation worksheet. Classify each gap as counting, timing, qualification, revenue, source, or assisted-path. Return owner, likely cause, and the next question to ask.

textFind decisions in this budget review that are unsafe because platform-reported performance does not match closed revenue. Explain what evidence is missing.

textTurn this reconciliation table into a one-page operator summary: what changed, what is safe to decide, what needs owner review, and what should not move yet.

Do not paste sensitive customer data into an unapproved tool. Use row IDs, aggregated tables, or approved internal AI environments when the data is private.

The monthly operating rhythm

Run this once per month before budget review.

1. Export system-of-record outcomes.
2. Export platform claims for the same period.
3. Normalize definitions.
4. Compare totals.
5. Match records where possible.
6. Classify the gap.
7. Assign owners.
8. Write the decision log.
9. Decide only what the evidence supports.

The first run may be messy. That is expected. Mess is the signal. The second run should be cleaner because the team knows which fields, conversion actions, and owner rules need repair.

What good looks like

A good reconciliation review ends with plain-language decisions:

• We can increase this campaign because closed revenue supports the platform trend.
• We cannot cut this channel yet because the sales cycle falls outside the platform window.
• We need to repair offline conversion uploads before judging nonbrand search.
• We need to split new and returning customers before trusting ROAS.
• We need finance to define whether this review uses gross revenue, net revenue, or margin.

That is the point. The business stops arguing about which dashboard is right and starts deciding what can safely change.