- Ad platform
What the platform reports under its attribution rules.
- Activity records
What your website, forms and call records captured.
- Business records
What your customer and accounting records confirm.
Check definitions, dates, attribution windows and duplicate records.
These systems can share data. Matching totals do not prove causation; different totals need an explanation, not an average.
Give each source a job
I use this practical review framework to give each source a clear job. The three sources answer different questions. I do not expect identical totals; I expect your team to explain the important differences. These systems may share data, so agreement does not prove a campaign caused the result.
| Source | What it tells your team | What to question |
|---|---|---|
| 1. Advertising platforms | What you spent, where ads ran and which actions the platform credits to advertising. | What counts as a conversion? Which attribution model and time window apply? |
| 2. Customer activity records | Recorded website actions, calls, forms, bookings and orders outside the ad dashboard. | Are events duplicated? Is tracking missing? Are repeat contacts counted as new opportunities? |
| 3. CRM and accounting | Which opportunities qualified, sold, cancelled, paid or received refunds. | Are stages updated consistently? Does reported revenue mean booked work, completed sales or cash collected? |
Compare the same thing before debating the result
Attribution means assigning credit to marketing interactions. A conversion window limits how long after an interaction a platform can credit an outcome. Changing the model or window can change reported performance without changing the customer’s actual purchase.
Keep the original exports. Then create a comparison using a consistent time zone, currency, date range and business stage. Check whether each report uses the date of the ad interaction, the inquiry or the sale. For a longer sales cycle, follow a cohort—a group of inquiries received during the same period—until enough time has passed for outcomes to develop.
Use a stable lead, opportunity or order ID to connect records. Keep repeat calls and forms attached to the relevant opportunity. Three contacts from one person do not automatically equal three new leads. Record the source exactly as received, alongside your cleaned reporting category, so a tidy dashboard does not erase the evidence.
Make tracking useful without claiming perfect visibility
UTM tags are labels added to campaign links. Use consistent source, medium and campaign names for email, social and other tagged links. Google Analytics uses these parameters to identify referring campaigns. A QR code on a mailer can open a tagged page; a dedicated phone number can help identify responses to a placement.
These methods improve coverage. They do not capture every influence. Someone may see your event booth, search for your brand later and call from another device. Ask how they heard about you, and retain that answer as additional context. It is a person’s recollection, not proof of which advertisement caused the sale.
Analytics also has gaps from consent choices, blocked tracking and implementation problems. Missing campaign tags do not prove a visit was organic. If you cannot establish the source, keep it unknown and investigate recurring gaps.
Know the difference between credit and impact
Two platforms may both claim credit for the same sale. Adding their reported revenue or ROAS together does not produce your business’s total return. Reconcile actual orders and revenue separately.
View-through reporting credits some outcomes after an ad impression without a click. Connected TV reporting may also use household matching; Google describes household-level measurement using IP addresses and other signals. Ask your provider which method, window and limitations apply. Those methods do not, by themselves, establish that the advertising caused additional sales.
Incrementality asks a different question: what happened because you advertised that would not otherwise have happened? A properly designed experiment with a comparison group helps answer it. Google explicitly distinguishes incremental conversions from standard attributed conversions. Your three-source check improves accountability; it does not replace that experiment.
Feed back outcomes your business can defend
Your CRM is the system where your team records customer relationships and sales progress. Agree on qualified and sold definitions there before sending outcomes back to advertising platforms. Google’s enhanced conversions for leads supports matching imported offline outcomes using customer data. It cannot fix an inaccurate sales status.
Have your specialist send only the approved outcomes and permitted data through the supported integration. Apply the relevant consent requirements and provider terms. Never put names, emails or phone numbers in tracking URLs; Google Analytics prohibits sending personally identifiable information. Review failed imports and corrections as part of the process.