Ad platform vs CRM reconciliation is the process of lining up the leads Google Ads or Meta reports with the records the client's CRM actually holds, then explaining every difference. The two will never match exactly, because they count different things on different dates with different rules. The goal is not zero gap; it is a gap you can break down into known causes, fix where tracking is broken, and explain where it isn't.
Key takeaways
- Most of the gap usually comes from a handful of causes: duplicates, spam, counting settings, attribution and date rules, modeled conversions, and leads lost between the form and the CRM.
- Totals can't tell you which cause you have. Matching records one by one, by click ID, email or timestamp, can.
- Some of the gap is legitimate and should stay: modeled conversions, view-through credit and different attribution rules are not errors.
- A simple reconciliation table, shared with the client, turns an accusation ("the ads are lying") into a list of fixes and known differences.
- Run it monthly. A gap that is explained every month stops being a crisis.
Why do Google Ads and the CRM report different lead counts?
Every agency gets this call eventually. The client opens their CRM, counts the leads from last month, and compares it with the ad platform report.
Example
Illustrative numbers. Google Ads reports 71 conversions for October on the campaign the client cares about. Their CRM shows 22 new contacts with Google Ads as the source for the same month. The client asks whether the ads are inflating results, and whether they've been paying for 49 leads that don't exist.
The honest answer is: probably not, but possibly partly. The ad platform and the CRM are measuring different things.
- The ad platform counts conversion events it attributes to ads, under its own counting settings, on the date of the click or impression, sometimes including modeled and view-through conversions.
- The CRM counts records, usually on the date they were created, with a source field that depends on how the form, the UTMs and the CRM's own attribution logic were set up.
Both can be correct and still disagree. Both can also be broken. Reconciliation is how you find out which parts of the gap are which.
What causes the gap between ad platforms and the CRM?
We sort every gap into the same categories. Some inflate the ad platform number, some deflate the CRM number, and some are just different rules.
Duplicates
One lead counted two or three times by the platform. Common causes: a thank-you page that fires the conversion on every reload, two tags for the same action (one hardcoded, one in GTM), a plugin and a GTM tag both sending the Meta Lead event, or Pixel and Conversions API events without a shared event ID. Our guides on duplicate Meta Pixel events on Shopify and WordPress and Meta Conversions API deduplication cover the Meta side in detail.
Spam and bots
Junk submissions trigger the conversion tag the same way real ones do. The CRM may filter them out (or the client's team deletes them), so they exist in the platform and nowhere else. Spam also teaches bidding to find more spam, which is a reason to fix it beyond reporting.
Counting settings
Google Ads lets each conversion action count "Every" conversion after a click or only "One." A lead action set to "Every" counts a person who submits two forms as two conversions. The CRM sees one contact. Google's help on conversion counting options recommends "One" for most lead actions.
Also check which actions are included. The Conversions column only includes primary actions, but "All conv." includes secondary ones. If the client's report includes page views, clicks on a phone number or chat starts, the platform number isn't a lead count at all.
Attribution windows, and conversion date vs click date
Google Ads attributes conversions to the date of the ad click, not the date of the conversion. A click on September 28 that converts on October 3 is a September conversion in Google Ads and an October record in the CRM. The "Conversions (by conv. time)" column reports by conversion date instead, which makes month-to-month comparisons with the CRM much fairer. Google's help on understanding conversion tracking data explains both views.
Windows matter too. Google counts conversions within the click-through window set on each action, which can be up to 90 days. Meta reports conversions under the attribution setting chosen for the ad set or the report, such as seven-day click or one-day view. Meta's reporting date logic also differs from a CRM's record creation date: depending on the report time setting, Meta can show a conversion on the date of the ad impression rather than the date the conversion happened. Make sure you know which window and which date each number uses before comparing anything.
Modeled conversions
Google models conversions it can't observe directly, for example because of consent choices or browser limits, and includes them in the Conversions column. Google describes this in its help on modeled online conversions. Modeled conversions are estimates of real conversions; they have no matching record by design. Meta also uses modeling in some reporting. Treat this part of the gap as legitimate, but don't let it become a bucket for unexplained numbers.
View-through and engaged-view conversions
Display, YouTube and Meta can credit a conversion to someone who saw an ad and converted later without clicking. The CRM's source field, built from UTMs or referrers, will usually say organic, direct or another channel for that person. Both are true. They are answering different questions.
Cross-device and cross-channel credit
A person clicks a Google ad on their phone, comes back on a laptop through a branded search or a direct visit, and submits the form. Google can credit the ad through signed-in data. The CRM, looking at the laptop session's UTMs, records direct or organic. The same lead can also be claimed by Google, Meta and the CRM's own "original source" at once. Platforms each claim the conversions they influenced; they don't split credit with each other.
Time zones
The ad account, the CRM and the export tool may each use a different time zone. Leads near midnight on the first and last day of the month land in different months. On small accounts, this alone can move several leads.
Leads lost between the form and the CRM
This is the category that matters most, because these are real leads the client never sees. The tag fires on submit, but the form's integration with the CRM fails: an API key expired, a required field was added in the CRM, a Zapier step errored, a booking tool sends to email but not to the CRM, or the form plugin stores entries in WordPress and nobody connected it. The platform is right, the CRM is wrong, and the client is losing business.
Sales team not logging
Phone leads that are never entered, leads created by hand without a source, existing contacts whose new enquiry updates an old record rather than creating a new one. The CRM undercounts, and nothing in tracking can fix it. It needs a process change, and it's worth naming politely in the write-up.
Diagram
Suggested diagram: a horizontal waterfall chart starting from the ad platform's conversion count on the left and stepping down through labelled bars (duplicates, spam, counting settings, date and window differences, modeled and view-through, cross-device credit) to the matched leads, with a separate upward bar for leads missing from the CRM, ending at the CRM count on the right. All numbers labelled as illustrative. Alt text: "Waterfall chart breaking an illustrative gap between ad platform conversions and CRM leads into its causes."
How does ad platform vs CRM reconciliation work, step by step?
Here's the method we use. It works for Google Ads and Meta, and for any CRM that can export records with timestamps.
- Fix the scope. Pick one client, one conversion action (or Meta event), one CRM pipeline and one closed date range, ideally a full month that ended at least a couple of weeks ago so late conversions have arrived.
- Align the definitions. Write down for each source: what is counted, which date is used, which time zone, which attribution window, and whether modeled or view-through conversions are included. For Google Ads, pull both Conversions and Conversions (by conv. time).
- Export the platform side at the most granular level available. For Google, conversions by day and campaign, plus any per-conversion data you have (form submission logs, GA4 events with timestamps, enhanced conversions or offline import records). For Meta, events by day plus the lead form export for Instant Forms. Note what's modeled where the platform shows it.
- Export the CRM side. All records created in the range, not only those marked as paid search, with created time, source fields, gclid, fbclid, UTM fields, email and any deletion or spam flags.
- Collect the website side. Form entries from the form tool, if it keeps them. This is the referee: it shows what was actually submitted, independent of both the platform and the CRM.
- Match record by record. Use the strongest key available: click ID first, then email or phone, then timestamp within a few minutes plus form name. Mark each row as matched, platform-only or CRM-only.
- Classify every unmatched row into the categories above: duplicate, spam, counting, date or window, modeled or view-through, cross-device or other-source credit, lost between form and CRM, not logged.
- Quantify what remains. Anything you can't classify is the true unexplained gap. Keep it as its own line; don't fold it into "modeled."
- Fix the broken parts (duplicates, spam counting, lost leads, missing click IDs) and test each fix with a real submission.
- Write the summary for the client in plain language, with the table below.
Record-level matching is much easier if click IDs are already stored in the CRM. If they aren't, adding them is the first fix: see our guide to lead source attribution in HubSpot and GoHighLevel.
