Conversion definition document
A written definition of primary and secondary conversions, agreed with your team before any tracking gets built.
Measurement
Conversion tracking built around what actually counts as a win
Bad conversion data does not just misreport performance. It actively teaches bidding algorithms to chase the wrong people. Getting this right is worth more than most creative or budget changes.
Every ad platform will happily let you count a form submission, a button click, a page view, and a phone number reveal as four separate conversions for the same lead, then optimize toward all four at once. Most accounts we inherit are doing exactly this without realizing it, which means the bidding algorithm is being trained on noise dressed up as signal.
Conversion tracking work starts with a conversation that has nothing to do with pixels: what is a real conversion for this business. For an ecommerce brand that might genuinely be a completed purchase. For a services business it is far more often a qualified lead that a human would recognize as worth pursuing, not every form fill regardless of quality. Getting that definition wrong upstream makes everything downstream wrong too, no matter how clean the technical implementation is.
Once the definition is settled, the technical work is deduplication and completeness. Deduplication means the same conversion is not counted twice across GA4, the ad platform's own pixel, and a CRM sync, which inflates reported performance and confuses attribution. Completeness means capturing what happens after the click: offline conversions like a phone call that closes, and CRM-side outcomes like a lead that later gets marked disqualified or a deal that closes weeks later, long after the ad platform stopped watching.
This work sits underneath the paid media accounts we run and the ones we do not. Bidding algorithms on every platform now lean heavily on conversion signal to decide who to show ads to. Feed them a clean, complete signal and they get better at finding the right people. Feed them noise and they get confidently wrong.
What it is
What the work actually involves
Definition, deduplication, offline conversion import, and CRM feedback loops.
We start by defining, in writing, what counts as a conversion for this business and why. This usually means separating a primary conversion (a qualified lead, a purchase, a booked call) from secondary engagement signals that are useful to watch but should not be optimized toward with the same weight.
Deduplication involves auditing every place a conversion is currently being counted: GA4, each ad platform's native pixel or conversion API, and any CRM or marketing automation sync. We identify overlap and configure primary conversion actions so bidding algorithms see one clean signal per real event, not three inflated ones.
Offline conversion import matters wherever the real outcome happens off the website: a phone call that gets booked and later closes, a walk-in, a quote that converts weeks after the original ad click. We build the pipeline, usually through a CRM export or API connection, that feeds those outcomes back to the ad platform tied to the original click or user identifier.
CRM-side conversions close the loop further. A form fill that turns out to be spam or a disqualified lead should not carry the same weight as one that becomes a customer. Where the CRM supports it, we build the sync that reports lead quality and deal outcomes back into the ad platform, so the algorithm eventually learns to prefer the traffic that actually closes.
Fit
Who this is for, and who it is not for
We would rather say no early than sell a program that cannot work.
Right fit
- You are running paid media and suspect the reported conversion numbers do not match what sales or fulfillment sees.
- A meaningful share of your real conversions happen offline or on the phone, and none of that is currently being fed back to ad platforms.
- You want bidding algorithms optimizing toward lead quality, not just lead volume.
Not the right fit
- You have no paid media running and no CRM to close the loop with. Start with GA4 setup instead.
- You want a single vanity number to report upward regardless of whether it reflects real business outcomes.
- Your sales cycle and CRM data are too unstructured to support any kind of offline conversion feedback yet.
Deliverables
What you get
A conversion setup that reflects real business outcomes, not just clicks.
Deduplication audit
Every current conversion source mapped and reconciled so the same event is not counted multiple times across platforms.
Offline conversion import
A pipeline that feeds phone, in-person, or delayed outcomes back to the ad platform tied to the original interaction.
CRM feedback loop
Lead quality and deal outcomes synced back to ad platforms where the CRM supports it, closing the loop past the initial form fill.
Primary conversion configuration
Ad platform conversion actions configured and weighted so bidding algorithms optimize toward the outcomes that matter.
Signal quality report
A before-and-after view showing how many previously counted 'conversions' were duplicates, low quality, or not real conversions at all.
How we run it
How the work runs
Definition first, technical implementation second.
Step 1: Define the real conversion
We work with your team, including sales if relevant, to agree on what a genuine conversion looks like for this business.
Step 2: Audit current tracking
We map every place a conversion is currently counted and identify duplication and gaps against the agreed definition.
Step 3: Build deduplicated tracking
Primary conversion actions get configured cleanly across GA4 and each ad platform, with duplicates removed or downgraded.
Step 4: Connect offline and CRM data
Where offline outcomes matter, we build the import pipeline and CRM sync needed to feed them back to the platforms.
Step 5: Validate against real outcomes
We compare reported conversions against actual sales or fulfillment data to confirm the numbers now agree.
Why does bad conversion tracking hurt ad performance?
Bidding algorithms use conversion signal to decide who to target. Duplicate or low-quality conversions train them to chase people who look like poor leads, actively worsening targeting over time rather than just misreporting results.
What is an offline conversion in this context?
An offline conversion is a real outcome that happens away from the website, such as a phone call that closes or a quote accepted weeks later, imported back to the ad platform so it can be credited to the original click.
Where this connects
Where this connects
Conversion tracking depends on clean measurement and feeds directly into paid media performance.
None of this works without a properly instrumented property, which is why it starts from GA4 setup and the event taxonomy built there.
Clean conversion signal is what makes reporting dashboards trustworthy instead of just precise-looking.
The ongoing tuning of what gets optimized toward is covered under strategy and optimization once the initial definitions are in place.
This work has the most direct impact on Google Ads since Google's automated bidding strategies rely almost entirely on the conversion signal we clean up here to decide who to bid on and how much.
Whether the improvements are actually moving the business is answered separately under success metrics which looks past raw conversion counts.
Questions
Conversion tracking questions we get asked
Stop optimizing toward the wrong conversions
Get conversion tracking that reflects what your business actually calls a win.