Why conversion numbers often do not match
A senior team reviews marketing performance.
Google Ads reports 48 conversions.
Meta reports 36 conversions.
GA4 shows 57 form submissions.
The CRM shows 41 new leads.
Sales says only 19 were qualified.
None of these numbers are automatically wrong. They are just answering different questions.
| System | What it usually shows | What it does not always prove |
|---|---|---|
| Google Ads | Conversions attributed to Google activity | Whether the lead became qualified or sold |
| Meta Ads | Conversions linked to Meta views or clicks | Whether Meta created demand or assisted it |
| GA4 | Website events such as forms, calls or purchases | Whether the event became a real business outcome |
| CRM | Leads, opportunities and sales stages | Whether the original marketing source was captured correctly |
| Sales / finance | Qualified leads, closed sales or revenue | Which touchpoints influenced the journey |
This is why conversion reporting can become uncomfortable at senior level. The business is not only asking,
“How many conversions did we get?” It is asking:
Which number should we trust when making budget decisions?
A platform conversion can be useful for optimisation. A business conversion is different. It needs to connect the
enquiry, sale or revenue back to an agreed definition.
Without that source of truth, marketing performance may look precise while still being unclear.
The problem: each platform counts conversions differently
Marketing platforms are not designed to agree with each other. They are designed to measure performance from their
own view of the customer journey.
That creates a common reporting issue: the same customer action can appear in several places, under several rules,
with several different totals.
| Why numbers differ | Example | What this can do to reporting |
|---|---|---|
| Different attribution windows | Google counts a lead within 30 days of a click. Meta may count after a click or view within its own window. | Two platforms may claim influence over the same lead. |
| Different conversion definitions | One report counts a form submit. Another counts a phone click. The CRM counts only new enquiries. | The business compares numbers that are not measuring the same action. |
| View-through attribution | A person sees an ad, does not click, but later converts. | A platform may report a conversion even when the visible journey looks direct or organic elsewhere. |
| Duplicate event firing | A form event fires twice after one submission. | Conversion volume looks stronger than reality. |
| Missing deduplication | A lead submits twice, or the same enquiry appears from two tracking paths. | One person may become multiple reported conversions. |
| CRM source gaps | The lead enters the CRM, but campaign source, medium or keyword is missing. | Sales can see the lead, but marketing cannot confidently connect it to spend. |
| Offline sales delay | A lead converts today but becomes a sale weeks later. | Campaign reports look complete before the business outcome is known. |
This is not a small technical detail. It can change how performance is interpreted.
A campaign with 100 reported conversions may become:
| Reporting layer | Reported number |
|---|---|
| Platform-reported conversions | 100 |
| Website form submissions after removing duplicates | 82 |
| CRM leads created | 67 |
| Sales-qualified leads | 31 |
| Closed sales | 9 |
The issue is not that the first number is useless. It may still help optimise media delivery. But it should not be
treated as the final business result.
This is where many marketing reports lose senior confidence. They show performance, but not always the chain from
spend to qualified lead, sale or revenue.
Recent marketing analytics commentary points to the same problem at a wider level: data quality is seen as critical
by most leaders, yet confidence in current measurement remains much lower. The gap is not usually caused by a lack
of reports. It is caused by unclear definitions, disconnected systems and weak governance around what counts as the
real conversion.
Platform conversions vs business conversions
A conversion is not always the same thing as a business result.
A platform may count a conversion when it sees a tracked action after an ad click or view. That is useful for
campaign optimisation. But senior decisions usually need a different layer: whether that action became a qualified
lead, sale, booking or revenue.
| Conversion type | Example | Main use | Risk if used alone |
|---|---|---|---|
| Platform conversion | Google Ads or Meta reports a lead | Optimising campaigns inside the platform | May over-credit the platform or include assisted actions |
| Website conversion | GA4 records a form submit, call click or purchase event | Understanding website behaviour | May include duplicate, test or low-quality actions |
| CRM conversion | A lead is created, qualified or moved to opportunity | Measuring lead quality and sales progress | Source data may be missing or inconsistent |
| Sales conversion | A deal is closed, booking is confirmed or invoice is paid | Measuring business outcome | May not show which marketing activity influenced the sale |
| Revenue conversion | Actual revenue, customer value or repeat purchase | Budget and commercial decision-making | Often delayed and disconnected from campaign reporting |
This is why platform conversions and business conversions should not be treated as interchangeable.
A campaign can look strong in the ad platform and still produce low-quality leads. Another campaign may look modest
in clicks or conversions, but produce better sales outcomes. Without a clear link between campaign activity and
business results, the report may reward volume instead of value.
A useful way to separate the two is:
| Question | Best place to answer it |
|---|---|
| Which ads, audiences or keywords are generating tracked actions? | Ad platforms |
| What happened on the website before conversion? | GA4 / website analytics |
| Which enquiries became qualified leads? | CRM |
| Which leads became sales? | Sales system / CRM |
| Which marketing spend contributed to revenue? | Agreed reporting source of truth |
The goal is not to remove platform conversions. They are important. The goal is to stop using them as the final
answer when the business question is really about leads, sales or revenue.
Why businesses need one source of truth
One source of truth does not mean one tool replaces every other tool.
It means the business agrees which system is used as the final reference point for conversion reporting, and which
systems are used for supporting analysis.
For example:
| Business model | Practical source of truth | Why |
|---|---|---|
| Lead generation | CRM | It shows whether enquiries became qualified leads or sales opportunities |
| Ecommerce | Ecommerce platform / transaction data | It records actual purchases and revenue |
| Appointment-based business | Booking system or CRM | It confirms whether enquiries became real appointments |
| B2B or long sales cycle | CRM plus sales pipeline stages | It connects marketing activity to opportunity quality over time |
| Offline sales | CRM, sales system or finance data | It captures outcomes that ad platforms cannot fully see |
The source of truth should be close to the real business outcome. For many businesses, that is not the ad platform.
It is the system where the lead, sale or revenue is actually confirmed.
This creates clearer rules for reporting:
| Reporting question | Should it use platform data? | Should it use source of truth data? |
|---|---|---|
| Which campaign should we optimise today? | Yes | Sometimes |
| Which channel generated the most tracked enquiries? | Yes | Yes |
| Which channel generated qualified leads? | No, not alone | Yes |
| Which marketing activity created sales? | No, not alone | Yes |
| Where should next month’s budget go? | Partly | Yes |
Without this agreement, every report can become a negotiation.
Google Ads may show one version of success. Meta may show another. GA4 may show a third. Sales may see something
different again. The senior team is then left comparing numbers that were never designed to match.
With one source of truth, the conversation changes.
Instead of asking:
“Why are the numbers different?”
The business can ask:
“What do the differences tell us?”
That is a much better question. It allows platform data to support optimisation, while business data remains the
reference point for commercial decisions.
Example: one customer journey, multiple reported conversions
One customer journey can create several reported conversions.
A person may first click a Google ad, return later from organic search, see a Meta retargeting ad,
and then submit a form directly on the website.
In the report, this can look like several separate wins.
| Step | Customer action | What may be recorded |
|---|---|---|
| Monday | Clicks a Google Search ad | Google Ads records a click and may later claim the conversion |
| Tuesday | Visits again from organic search | GA4 records another session from organic |
| Wednesday | Sees a Meta retargeting ad | Meta may count a view-through or assisted conversion |
| Thursday | Submits a website form directly | GA4 records a form submit |
| Friday | Lead appears in CRM | CRM records one new lead |
| Two weeks later | Sales qualifies the lead | CRM records one qualified opportunity |
In platform reporting, this journey may appear as:
| System | Reported result |
|---|---|
| Google Ads | 1 conversion |
| Meta Ads | 1 conversion |
| GA4 | 1 form submission |
| CRM | 1 lead |
| Sales pipeline | 1 qualified opportunity |
This does not mean the business received five separate outcomes.
It may be one person, one enquiry and one qualified opportunity, reported through different systems.
This is why adding platform conversions together can be dangerous. Google Ads and Meta totals are useful inside
their own platforms, but they are not always additive at business level.
A simple example:
| Report view | Number shown |
|---|---|
| Google Ads conversions | 38 |
| Meta Ads conversions | 27 |
| LinkedIn Ads conversions | 12 |
| Total if added together | 77 |
| CRM leads created | 49 |
| Sales-qualified leads | 18 |
The total platform number may be much higher than the actual number of leads because several platforms may be
claiming influence over the same people.
The better question is not only:
Which platform reported the conversion?
The better question is:
Which conversions became real business outcomes, and what role did each channel play?
What should and should not be treated as the source of truth
The source of truth should not be chosen because a system has the biggest number, the cleanest dashboard or the most
familiar interface.
It should be chosen based on two business rules:
| Rule | Question it answers |
|---|---|
| Conversion source of truth | Where do we confirm that a real lead, sale or revenue event happened? |
| Attribution source of truth | How do we decide which channel, campaign or touchpoint gets credit? |
Both rules matter.
A CRM may confirm that a sale happened. But unless the business also agrees how attribution works, the sale can still
be interpreted in different ways.
For example:
| Journey | Possible attribution rule | Channel credited |
|---|---|---|
| User first clicks Google Ads, returns later direct and buys | First-touch attribution | Google Ads |
| User first finds the website through organic search, later clicks Meta and converts | Last-click attribution | Meta |
| User clicks Google Ads, then email, then submits a form | Last non-direct click | |
| User sees Meta, clicks Google Ads, later becomes a sale | Multi-touch / assisted view | Meta and Google both receive partial credit |
| User comes through Google Ads but sale is closed offline 3 weeks later | CRM-based attribution | Google Ads, if original source was captured correctly |
This is why “agreed” does not just mean everyone looks at the same CRM number.
It means the business agrees:
| Attribution decision | Why it matters |
|---|---|
| First touch, last touch or multi-touch? | Changes which channel appears to create value |
| How long is the attribution window? | A 7-day and 30-day window can produce different results |
| Should direct traffic get credit? | Direct is often a return visit, not the original source |
| How are view-through conversions treated? | Important for Meta/display, but can inflate results if not controlled |
| How are offline sales connected back? | Critical for long sales cycles and lead generation |
| What happens when source data is missing? | Prevents “unknown” or “direct” from hiding marketing impact |
A practical reporting setup gives each system a clear role:
| System | Role in reporting | Should it be final authority? |
|---|---|---|
| Ad platforms | Campaign optimisation and media diagnostics | No |
| GA4 / website analytics | Website behaviour and journey analysis | Not usually |
| GTM / tracking setup | Event collection and technical logic | No, but it controls data quality |
| CRM | Lead record, qualification and pipeline progress | Often yes for leads |
| Ecommerce / booking system | Confirmed transaction or booking | Often yes for transactions |
| Finance / sales system | Revenue and commercial outcome | Yes for revenue |
| Attribution logic / agreed model | Decides how credit is assigned to channels | Yes for channel performance decisions |
This avoids the common mistake of forcing every number to match. They do not need to match perfectly. They need to be
understood correctly.
For example, if Meta shows 120 conversions and the CRM shows 76 leads, the next step should not be to ask which
system is “wrong”. The better question is:
What exactly did Meta count, what exactly did the CRM accept, and what attribution rule are we using?
That question usually reveals the real issue: attribution window, duplicate events, lead rejection, missing source
data, form errors, offline delay or sales qualification rules.
The source of truth is not just a reporting preference. It is a business rule.
The goal is not to reduce the amount of data. It is to make the hierarchy clear.
Platform data explains how marketing activity performed.
CRM, sales or revenue data confirms what happened.
Attribution rules explain which marketing activity gets credit.
