Understand the numbers.
Connect them to outcomes.
Campaign reports, website analytics and your CRM each tell part of the story. KOAT helps configure measurement, investigate discrepancies and connect recorded actions with the results your business cares about.
Why do advertising results differ from CRM results?
The systems may count different actions, use different attribution rules or place results in different reporting periods. A platform may report a form submission while your CRM reports unique qualified leads or closed deals. Consent, missing identifiers, duplicate events and failed integrations can create further differences. First align the definitions, then investigate the measurement journey.
What exactly are you counting?
Someone takes a step
A button click, form start or telephone-link click indicates an interaction. It does not establish that a valid enquiry or conversation occurred.
A request reaches your business
A successfully received submission or booking is a stronger milestone. Check that the information reaches the agreed lead system.
The enquiry progresses
Qualification, attended appointments and closed deals describe later stages. These need their own definitions, timestamps and reliable records.
A “conversion” needs a definition.
State the action, counting rule and source behind the number. Ten tracked clicks, ten form submissions and ten new customers represent different business events.
Common reasons the totals differ.
| Possible cause | How it affects the comparison | What to check |
|---|---|---|
| Different funnel stages | The platform counts submissions; the CRM report counts qualified leads or sales. | The event and CRM stage behind each total. |
| Different counting units | Several submissions may become one CRM contact, while an existing contact may make a new enquiry. | Events, unique people, enquiries and opportunities. |
| Attribution rules | Systems may credit different touchpoints or use different attribution windows. | Model, eligible interactions, window and report scope. |
| Dates and time zones | An interaction, submission and closed sale can fall in different reporting periods. | Date basis, account time zones and conversion delay. |
| Measurement coverage | Consent, browser restrictions or missing identifiers can limit observed or attributable activity. | Consent behaviour, tag delivery and source availability. |
| Duplicate measurement | Multiple implementations or repeated triggers can record one action more than once. | Plugins, tags, imports and browser/server overlap. |
| Integration failures | The event is recorded, but the lead or its source information does not reach the CRM correctly. | Delivery logs, field mapping, routing and permissions. |
| Processing and exclusions | Delayed reporting, spam filtering or removed test records can change totals. | Processing time, filters and record history. |
Some differences are expected. Others reveal a fault.
Document the explanation for each material gap. Different attribution rules may explain a reporting difference; a failed lead delivery needs a repair. There is no universal percentage that makes every discrepancy acceptable.
Build a measurement setup with a purpose.
Understand website activity
Configure relevant events and reports for the customer journey. Important actions can be marked as key events; a Google Ads conversion can be created from a suitable Analytics key event.
Manage tag implementation
Organise tags, triggers and variables around a documented event plan. Preview and debugging tools help check a configuration before publishing it.
Support Meta event measurement
Configure the agreed website events and review whether they represent the intended actions. Check existing integrations to avoid overlapping implementations.
Your CRM records the sales process.
Use it to record enquiries, ownership, qualification and outcomes. Its source field also needs scrutiny: a manually entered label or overwritten campaign field can produce misleading channel reports.
Track the completed action.
A submit-button click can happen before validation succeeds. A thank-you page can sometimes be revisited. Design the event around a reliable confirmation that the intended action occurred.
Keep early interactions separate from completed enquiries. This helps reports describe the journey and keeps campaign goals aligned with meaningful actions.
Document every important event
- The business action it represents.
- The condition that triggers it.
- The system responsible for sending it.
- The required parameters and value definition.
- The intended counting and duplicate handling.
- The reports and campaign goals that use it.
A telephone or WhatsApp click should be labelled as a click unless another reliable system confirms the resulting conversation.
Compare like with like.
Imagine an advertising report shows 40 attributed lead events. Your CRM contains 32 unique contacts, 18 qualified enquiries and five closed deals. These hypothetical numbers do not, by themselves, establish a tracking fault.
| Reported number | Question to resolve |
|---|---|
| 40 attributed lead events | Were these successful submissions? What counting, attribution and date rules were used? |
| 32 unique CRM contacts | Were repeat submissions merged? Did some enquiries come from contacts already in the CRM? |
| 18 qualified enquiries | What qualification criteria were applied, and have all relevant enquiries been assessed? |
| Five closed deals | Do they belong to the same enquiry cohort, and has enough time passed for the sales cycle? |
Use enquiry cohorts for sales progression.
Follow enquiries created in a defined period into their later outcomes. Comparing this month’s submissions with this month’s closed deals can mix new enquiries with sales from earlier periods.
Test the journey from action to record.
A valid action is recorded
Complete the intended enquiry or purchase and check the event, relevant values and resulting business record.
An unsuccessful attempt stays separate
Test validation errors and failed submissions. Confirm they do not masquerade as completed enquiries or purchases.
Repeated actions are understood
Check reloads, repeat submissions and overlapping implementations against the agreed counting rules.
Choices affect tags as intended
Test the relevant consent states and transitions. Verify the implementation reflects the agreed consent configuration.
The lead reaches its destination
Confirm field mapping, campaign information, assignment and notifications in the CRM or agreed lead system.
The published version is checked
Verify the live configuration after release and document the result. Recheck affected paths after material website, form or integration changes.
Consent and data handling belong in the setup.
Configure measurement around the business’s consent requirements and relevant platform policies. Google’s consent mode adjusts Google tag behaviour according to consent state; the implementation still needs testing.
Keep names, email addresses, phone numbers and form message text out of ordinary GA4 event parameters and page URLs. Review what tags actually transmit.
Server-side tracking has limits too.
A server-side or Conversions API setup may be part of the agreed scope. It still requires appropriate data handling, event definitions and duplicate controls where the same action is also sent through a browser.
It does not guarantee complete measurement or permission to disregard a person’s choices.
Give each report a clear job.
Use platform reports to understand attributed campaign performance under their stated rules. Use GA4 to examine measured website journeys and acquisition. Use verified CRM and order records to assess operational outcomes.
Bring these views together with documented definitions. Where information is missing or a result is modelled, explain the limit instead of presenting false precision.
Attribution is different from causation.
Several platforms may claim credit for the same sale. Do not add their attributed conversions as though every total represents separate customers.
Tracking shows recorded or attributed activity. Assessing the additional outcomes caused by advertising may require a suitable experiment.
Define. Implement. Validate. Explain.
- 01
Map the measurement
Review your website, campaigns, CRM and existing implementations. Agree on meaningful events and outcomes.
- 02
Configure the setup
Implement or repair the agreed tags, events and integration requirements.
- 03
Test the complete path
Validate actions, parameters, consent behaviour and business records across the relevant scenarios.
- 04
Document the result
Explain definitions, verified behaviour, known limits and the checks needed after future changes.
What does conversion tracking setup cost?
Cost depends on the website, event requirements, existing tags, consent configuration and integrations. A simple enquiry form differs from a multi-domain purchase journey or CRM outcome connection.
We review the setup before defining the work. Advanced server-side implementation, dashboards and ongoing monitoring can be included where relevant to the agreed scope.
Bring these to the first conversation
- Your website and important enquiry or purchase paths.
- Relevant GA4, GTM and advertising account access.
- Your CRM and lead delivery process.
- Your consent management setup.
- The reports or totals that appear inconsistent.
- Your definitions of a lead, qualified enquiry and sale.
Tracking and measurement explained.
Why do advertising and CRM totals differ?
Check the event, counting unit, reporting dates and attribution rules first. Then investigate duplicate measurement, missing source information, consent coverage and lead delivery. Qualification and closed sales are later stages than submission.
Should all systems show identical numbers?
No. They can serve different purposes and apply different rules. The goal is to understand material differences and repair faults, while making valid comparisons between equivalent measures.
What is the difference between GA4 and Google Tag Manager?
GA4 collects and reports measured activity. Google Tag Manager manages the tags and conditions used to send data to tools. Installing a GTM container alone does not create a complete event measurement plan.
What are GA4 key events?
They are events marked as important to your business. A suitable GA4 key event can also be used to create a Google Ads conversion. Review the conversion configuration and campaign goal settings separately.
Does a form-button click count as a lead?
It records a click unless the setup verifies a successful enquiry. Someone may click before validation succeeds. Use a reliable completion condition for the lead event.
Can a WhatsApp or phone click be tracked?
Yes, as an interaction where technically measurable. The click does not confirm a completed conversation. Measuring that outcome requires a suitable additional process or system.
Why are conversions appearing twice?
Possible causes include duplicate tags, overlapping plugins, repeated confirmation triggers or several conversion actions counting the same business outcome. Review the whole implementation before removing anything.
Can CRM outcomes inform advertising measurement?
Where supported, a suitable integration can send defined outcomes such as qualified leads or sales to an advertising platform. Requirements depend on the platform, identifiers, permissions and data quality.
Does server-side tracking capture everything?
No. Coverage still depends on the implementation, available information, consent and platform processing. It also needs checks for duplicate events and accurate values.
When should tracking be checked again?
Recheck relevant paths after changes to forms, checkout, plugins, consent tools, domains or integrations. Unexpected changes in event volume or differences with business records also justify investigation.
Know what is counted.
Understand what it means.
Let’s review your tracking setup and reporting differences, then define the work needed to connect measurement with meaningful business outcomes.
