Why Your Google Ads Conversion Numbers Don’t Add Up (And How to Audit Conversion Tracking)

If you’ve ever pulled up your Google Ads dashboard, then opened GA4 or your CRM in another tab, and found three different conversion counts for the same week — you’re not doing anything wrong. That mismatch is normal. What isn’t normal is not knowing why the numbers disagree, because that gap is often hiding the exact problem that’s quietly wrecking your Smart Bidding.

Conversion tracking is the one part of a Google Ads account where a small, invisible error compounds. A broken search term list wastes money on the wrong queries. A stale ad wastes attention. But broken conversion tracking corrupts the signal Google’s bidding algorithms use to decide who to show your ads to, how much to bid, and which campaigns to favor with budget. Get it wrong for a few weeks and you’re not just misreporting performance — you’re actively training Smart Bidding to chase the wrong outcome.

This post is a practical walkthrough of why conversion numbers drift, the specific ways tracking breaks in real accounts, and a step-by-step audit you can run this week without needing a developer.

The Number Everything Else Depends On

Every optimization decision in a modern Google Ads account — budget allocation, Target CPA, Target ROAS, Maximize Conversions, even which keywords Performance Max leans into — is downstream of one thing: what you told Google counts as a conversion, and how faithfully that signal is being recorded.

If your “conversion” is actually a page view that fires whether or not someone submitted a form, Smart Bidding will happily spend your budget getting more page views. If a tracking tag stopped firing three weeks ago after a site update, the algorithm isn’t seeing “we stopped converting” — it’s seeing “conversions collapsed,” and it will often react by pulling back reach or shifting spend to the campaigns that still look like they’re working, whether or not that’s actually true.

This is why conversion tracking deserves to be the first thing checked in any account review, before keyword lists, before ad copy, before bid strategy. Everything else is optimizing toward a target — and if the target is wrong, faster optimization just gets you to the wrong place quicker.

Why Your Numbers Will Never Perfectly Match — And What Gap Is Normal

Before auditing anything, it helps to accept a baseline fact: Google Ads, GA4, and your CRM are not measuring the same thing, and they were never going to agree exactly. Each one answers a different question. Google Ads asks which ad interaction should get credit for a conversion, for the purposes of bidding and reporting. GA4 asks what happened on your website or app, and which channels contributed along the way. Your CRM asks which of these are actually usable leads, opportunities, or paying customers.

Some of the most common — and entirely normal — reasons the numbers don’t line up:

• Attribution differences. Google Ads will credit a sale to a Google click if that click appears anywhere in the path, depending on your attribution model. Your CRM might only credit the very last touch. That alone can make Google Ads report more “Google-driven” conversions than your CRM agrees with.
• Conversion counting settings. A conversion action set to count “every” conversion instead of “one per click” means a single click followed by three purchases inside the conversion window reports as three conversions, not one. That’s not a bug — it’s a setting — but it’s frequently the wrong setting for lead-gen accounts.
• Consent and modeled data. When visitors decline cookies, GA4 may not record their session at all, while Google Ads can still model some of that activity through Enhanced Conversions or Consent Mode. Your CRM, meanwhile, logs the sale regardless of what any cookie banner allowed.
• Offline and delayed conversions. Phone calls, walk-ins, and sales that close weeks after the click often exist in the CRM long before — or without ever — showing up in GA4 or Ads, unless you’re importing offline conversions back in.
• Bot and spam traffic. GA4 tends to count form-page events that never turn into a real, human-reviewed lead in the CRM.

None of this means the mismatch should be ignored — it means the goal of an audit isn’t “make all three systems agree exactly.” It’s “understand the gap well enough to know when a new gap means something actually broke.”

How Conversion Tracking Actually Breaks

Tracking rarely fails all at once with an obvious error message. It degrades quietly, usually triggered by a change that had nothing to do with marketing. The most common causes worth knowing before you go looking:

• Site redesigns and theme changes. A new checkout flow, a redesigned thank-you page, or a new page builder can shift exactly where and when a conversion tag is supposed to fire — without anyone remembering to update the tag.
• Google Tag Manager triggers tied to URLs. If a conversion trigger only fires on a specific thank-you page URL, and that URL changes during a redesign, the trigger silently stops matching. No errors, no warnings — the tag just never fires again.
• “All Pages” triggers gone wrong. The opposite problem: a trigger scoped too broadly fires on every page load, inflating conversions that never happened. This is one of the more common causes of a sudden, unexplained conversion spike.
• Event naming drift. When a developer renames a dataLayer event during a frontend update without telling anyone, any GTM trigger depending on the old event name goes dark.
• Consent Mode misconfiguration. A consent banner update that defaults to blocking tags for all users — not just the ones who actually decline — can quietly cut off a large share of your tracked conversions.
• Duplicate tags. Two versions of the same conversion tag firing on the same page (often left over from a previous developer or a migration) will double-count conversions in a way that looks like “great news” until someone checks the CRM.

What ties these together is that none of them show up as an error in the Google Ads interface. The campaign just starts converting less, or more, and it’s easy to read that as a performance problem rather than a plumbing problem.

What “One Per Click” vs. “Every” Actually Does to Your Numbers

This setting is worth calling out on its own because it’s set once, buried in the conversion action’s settings, and then forgotten — even though it directly shapes every reported number afterward.

“One per click” counts a maximum of one conversion for each ad click, no matter how many times the person converts afterward. This is the right setting for most lead generation goals, where a second form fill from the same person isn’t a second lead. “Every” counts each individual conversion event, which is appropriate for ecommerce, where three separate purchases genuinely are three separate transactions.

The audit question is simple: does the counting setting on each conversion action actually match what that action represents? An account with “every” selected on a lead form is going to overstate lead volume, and Smart Bidding will optimize toward whatever is inflating the count — including, sometimes, toward the exact kind of low-quality repeat submission that shouldn’t be counted at all.

Enhanced Conversions: What It Fixes, and What It Doesn’t

Enhanced Conversions works by taking first-party data collected on your own site — a hashed email address or phone number, for example — and sending it alongside your existing conversion tag so Google can match conversions that happened across devices or after cookie loss (someone clicks an ad on their phone, then buys later on a laptop, for instance). It’s a genuinely useful layer on top of standard tracking, and it tends to recover conversions that would otherwise go unattributed as third-party cookies and cross-device tracking keep getting harder.

What it doesn’t do is fix a tag that isn’t firing at all, or correct a trigger that’s scoped to the wrong page. Enhanced Conversions improves the accuracy and completeness of conversions your tag actually captures — it has nothing to fix if the tag never captures the event in the first place. Checking whether Enhanced Conversions is set up and reporting correctly is a legitimate, worthwhile audit item; it’s just a different audit item from “is the base conversion tag firing.”

A Conversion Tracking Audit You Can Run This Week

None of this requires a developer for the first pass. Here’s a practical order to work through:

• 1. List every active conversion action and ask what it actually represents. Purchase, qualified lead, booked call — good. Page view, button click, or “visited pricing page” being used as a primary conversion for bidding — that’s a signal problem waiting to happen, even if the tag itself is working perfectly.
• 2. Check the counting setting on each action. Confirm “one per click” vs. “every” matches the actual business event, per action.
• 3. Compare Google Ads conversion volume against your CRM for the same date range. Don’t expect an exact match — expect a consistent, explainable ratio. If Google Ads reports 40 leads and your CRM shows 12 actual form submissions for the same period, that gap is worth chasing down, not shrugging off.
• 4. Trigger a real test conversion and watch it land. Submit your own form or complete a test purchase, then confirm it shows up correctly in Google Ads Tag Assistant (or GTM’s preview mode) and appears in the conversion action within the expected time frame.
• 5. Check the conversion window settings. A 30-day click window on a product with a 90-day sales cycle will systematically undercount, making the account look worse than it is right when Smart Bidding needs accurate signals most.
• 6. Verify Enhanced Conversions status. In the conversion action’s diagnostics, check whether Enhanced Conversions is active and whether match rate looks reasonable, rather than assuming it’s working because it was set up once.
• 7. Check consent mode behavior for a user who declines cookies. Open the site in a private window, decline the consent banner, and confirm tags behave as intended — blocking what should be blocked, not silently blocking everything.
• 8. Cross-reference any recent site or GTM changes with the account’s conversion trend line. If conversions dropped or spiked on a specific date, check GTM’s version history and any site deployment logs around that date before assuming it’s a market or seasonality issue.
• 9. Look for duplicate tags. View the page source or use GTM preview mode on your key conversion pages to confirm the conversion tag isn’t firing twice.

Working through this list takes an afternoon, not a sprint. The value isn’t finding a problem every time — it’s having a documented baseline, so the next time performance shifts, “check tracking first” is a five-minute step instead of a two-week mystery.

What to Do Once You Find a Problem

When the audit turns something up, resist the temptation to just reset the conversion action and move on. A few things matter more than speed here:

Fix the root cause before touching anything else — a mis-scoped GTM trigger, a renamed dataLayer event, a consent default — rather than layering a workaround on top of a tag that’s fundamentally pointed at the wrong event. If a conversion action has been broken or inflated for a while, be honest with yourself about how long Smart Bidding has been optimizing toward bad data, because bid strategies that rely on machine learning need time to recover once the signal is corrected — don’t expect performance to snap back the day you fix the tag. And document what changed and when, so if performance dips right after the fix, you can tell the difference between “the tracking is now accurate and the real numbers are lower than we thought” and “something else broke.”

Why This Works Better as a Recurring Check Than a Quarterly Fire Drill

Most of the tracking failures described above don’t announce themselves. They start on the day of a site update or a GTM publish, and they keep quietly distorting the account’s numbers until someone happens to compare Google Ads against the CRM and notices the gap has gotten worse. By the time a quarterly audit catches it, Smart Bidding has often had weeks to optimize toward a broken signal. This is a big part of what continuous account monitoring is actually for — not replacing a proper audit, but catching the day a conversion count moves in a way that doesn’t match the rest of the account, long before a quarter has gone by. Growera watches Google Ads accounts daily rather than on a scheduled review cycle, so a sudden drop or spike in conversions gets flagged close to when it happens — while it’s still a quick fix, not a rebuild of several weeks of bidding history.

The Bottom Line

Conversion tracking isn’t glamorous work, and it’s easy to treat it as something you set up once and leave alone. But it’s the foundation every other optimization in the account stands on. A perfectly tuned bid strategy pointed at the wrong signal will still spend your budget efficiently — just toward the wrong outcome.

The good news is that auditing it doesn’t require deep technical skill, just a willingness to actually compare what Google Ads is reporting against what your CRM knows to be true, and to treat unexplained gaps as something to investigate rather than something to average out. Run the checklist above once, fix what it turns up, and then make checking it a habit rather than a quarterly emergency. Your Smart Bidding strategy is only ever as good as the data it’s learning from.