When the Numbers Don't Add Up (Troubleshooting Guide)

Quick Answer: When marketing analytics numbers don't make sense, the cause is almost always one of six things: duplicate tags, broken cross-domain tracking, missing UTMs, conversion event misfiring, platform attribution model differences, or ad blocker impact. Work through them in order; the first four are technical fixes, the last two are interpretation issues.

Analytics breaks quietly. The dashboard still shows numbers, the reports still generate, but the numbers no longer reflect reality. A campaign launch doubles direct traffic instead of paid social. A site redesign drops conversion rate by 40 percent overnight. Google Ads reports 150 conversions, GA4 reports 90. None of these failures throw an error; they just produce data the owner has to either trust blindly or investigate.

The good news is that 90 percent of analytics problems trace back to six causes. Working through them in order resolves most issues in under 2 hours; complex cases needing developer intervention are the exception, not the rule. The frustrating news is that most owners don't know the diagnostic process and end up either ignoring the bad data (worst case) or rebuilding tracking from scratch (overkill in most cases).

This guide walks through the six most common breaks, how to diagnose each, and what to fix. Use it as a checklist when something looks wrong.

At a Glance

Quick Facts:

• Six causes of 90 percent of analytics problems: duplicate tags, cross-domain breaks, missing UTMs, conversion misfiring, platform attribution differences, ad blocker impact

• Diagnostic time for most issues: 30 minutes to 2 hours

• Free diagnostic tool: Google Tag Assistant Chrome extension

• Most common audit finding in Calgary SMB accounts: duplicate GA4 tag firing (inflates sessions by 2x)

• Hardest to diagnose: ad blocker impact (typically 10 to 30 percent data loss, invisible in reports)

• Easiest fix: missing UTMs (add and document for future)

Problem 1: Sessions Suddenly Doubled or Halved

Symptom: 

Site traffic jumps or drops 50 percent or more month-over-month without an obvious cause.

Most likely cause: 

Duplicate GA4 tag firing (doubled) or removed tag (halved). Common after site redesigns, theme updates, or developer changes.

Diagnostic: 

Install the Google Tag Assistant Chrome extension and load any page on your site. The extension shows every Google tag firing and how many times each fires. A single page should show GA4 firing exactly once. If it fires twice (or zero times), you have the problem.

Fix: 

For duplicate tags, identify which install method is firing both (usually one install via Google Tag Manager plus one hardcoded in the theme), and remove one. The cleanest setup keeps Tag Manager as the only install path. For removed tags, restore via Tag Manager and verify firing.

Verification: 

After the fix, Tag Assistant shows a single fire on multiple test pages. GA4 Realtime shows expected session counts within 60 seconds of test visits.

Problem 2: Direct Traffic Jumped Without Explanation

Symptom: 

Direct traffic share grows 20+ percent in a month. The growth doesn't come from a real direct traffic increase; it comes from other channels losing their source attribution and being recategorized as direct.

Most likely cause: 

Cross-domain tracking broken (visitors crossing from main site to subdomain or third-party form get reset) or UTMs missing from new campaigns.

Diagnostic: 

Check GA4 Reports > Acquisition > Traffic Acquisition for the affected month. If direct grew while another channel (typically paid social or email) shrank by a similar amount, the source attribution is being lost. Check recent campaign launches for missing UTMs on ad URLs and email links. Check booking and form subdomains for cross-domain tracking config.

Fix: 

Add UTMs to any campaign URLs that lack them (covered in cluster article 08). Configure cross-domain tracking in GA4 Admin > Data Streams > Configure Tag Settings > Configure Your Domains; add every subdomain and third-party domain visitors cross to.

Verification: 

Click a UTM-tagged campaign link and check GA4 Realtime; the source and medium should match the UTM values. Click through to a subdomain and verify the same session continues without resetting.

Problem 3: Conversions Stopped Firing or Doubled

Symptom: 

Conversion volume dropped to near-zero or doubled overnight. The change doesn't match actual business activity.

Most likely cause: 

Conversion event trigger broken (zero) or duplicate event firing (doubled). Common after form plugin updates, page URL changes, or new tags added that conflict with existing ones.

Diagnostic: 

Manually complete a test conversion (submit a form, click the phone link, complete a test purchase) and watch GA4 Realtime for the conversion event. If the event doesn't appear within 60 seconds, the trigger is broken. If the event appears twice, you have a duplicate firing.

Fix: 

For broken triggers, check Tag Manager: the trigger condition probably no longer matches the actual page URL or element selector. Update to match current site behaviour and republish. For duplicates, identify which two tags are firing (often the same event configured in both GTM and a platform-specific pixel) and remove one.

Verification: 

Repeat the test conversion and confirm single fire in Realtime. Check the conversion event count for the next 24 hours; it should match expected business volume.

Problem 4: GA4 and Google Ads Show Different Conversion Numbers

Symptom: 

Google Ads reports 200 conversions for the month; GA4 reports 130 for the same campaign. The discrepancy is 30+ percent.

Most likely cause: 

Different attribution models, different windows, or different conversion definitions. Often this is normal, not a bug.

Diagnostic: 

Compare the conversion definitions and attribution settings. Google Ads default is data-driven attribution with view-through credit; GA4 default may be the same model but a different window, or a different model entirely. Google Ads counts cross-device conversions natively; GA4 only does so when Google Signals is enabled.

Fix: 

Don't try to force the numbers to match across platforms; they measure different things. Pick GA4 as the source of truth for cross-channel analysis. For platform-level decisions (which Google Ads campaign to scale within Google Ads), trust the platform's numbers within its own context. Document the expected discrepancy range (typically 15 to 30 percent) so future reports don't trigger false alarms.

Verification: 

Run the comparison again next month; the percentage discrepancy should stay stable. A stable discrepancy is fine; a discrepancy that keeps growing signals a new problem.

Problem 5: Conversion Rate Dropped After a Site Change

Symptom: 

Site-wide conversion rate dropped 20+ percent after a site update, theme change, or new page launch.

Most likely cause: 

A CTA, form, or tracking element broke during the change. Less commonly, the change actually hurt UX (which the conversion data would correctly reflect).

Diagnostic: 

First, verify tracking. Test each conversion event manually; if any are firing inconsistently or not at all, the conversion rate drop is a tracking problem, not a UX problem. Second, check page-level conversion rates; if specific pages dropped much more than others, identify what changed on those pages.

Fix: 

Tracking-cause fixes are described in Problem 3 above. UX-cause fixes require CRO analysis: page speed, mobile responsiveness, form length, CTA visibility, copy changes. The discipline is fixing tracking first; never adjust pages based on conversion data you haven't verified is accurate.

Verification: 

Conversion rate returns to historical baseline within 14 days of fixing the underlying issue.

Problem 6: Real Traffic Doesn't Show Up in Reports

When the numbers don't add up, Calgary businesses should check whether GA4 is missing real traffic due to tracking limitations, using sales, calls, and customer feedback to identify data gaps.

Symptom: 

Owner knows the site got traffic (sales, calls, anecdotal evidence) but GA4 reports show much lower numbers than expected.

Most likely cause: 

Ad blockers, browser privacy settings (Safari ITP, Firefox tracking protection), and iOS app tracking restrictions are blocking the GA4 tag from firing for 10 to 30 percent of users. The data is gone; no fix recovers it.

Diagnostic: 

No clean diagnostic exists because the blocked sessions never appeared. Indirect signals include: server logs showing significantly more sessions than GA4, e-commerce platform sales counts higher than GA4 purchases by 20+ percent, anecdotal feedback that "the website is busier than the reports show."

Mitigation: 

Implement server-side tracking via Google Tag Manager Server-Side (technical, paid service typically $50 to $200/month) for businesses where the data loss meaningfully affects decisions. Use Conversions API for ad platforms to recover conversion attribution. Accept some data loss as the cost of operating in a privacy-conscious browser environment.

Verification: 

Server-side tracking implementation typically recovers 10 to 25 percent of previously blocked sessions. The gap closes but never reaches 100 percent of actual traffic.

Frequently Asked Questions

How do you tell if a number in your dashboard is wrong or just unexpected?

Three checks. First, does the number make sense given known business activity (campaign launches, sales results, anecdotal traffic)? Second, does the trend match year-over-year for the same month? Third, do related metrics tell a consistent story (high traffic should usually correlate with some conversion activity)? If all three pass, the number is probably right even if it's surprising. If any fail, investigate.

How often do analytics setups break on their own?

For Calgary SMBs, expect 2 to 4 break events per year on average. Site redesigns, plugin updates, ad campaign launches, and browser privacy updates all trigger breaks. The break itself is silent; only proactive auditing catches it quickly. Without auditing, breaks typically run 30 to 90 days before someone notices the data looks off.

Should you just rebuild GA4 from scratch when troubleshooting fails?

Rarely. Rebuilding loses all historical data and takes 15 to 25 hours. Reserve full rebuilds for cases where the existing setup is genuinely unsalvageable (which is unusual). For most issues, targeted fixes resolve the problem and preserve the data history.

Are there tools that monitor analytics health automatically?

Yes. Tools like Datatrue, ObservePoint, and AnalyticsCheckup automatically test tag firing, conversion events, and UTM consistency on a schedule. Pricing ranges from $50 to $500/month depending on coverage. Worth the investment for businesses spending $10,000+ per month on marketing where data accuracy is critical.

Does LTL Creative offer analytics audits and troubleshooting for Calgary businesses?

Yes. Analytics audits are available as a standalone engagement for businesses concerned about data accuracy, typically completed in 4 to 8 hours and including a written report of issues found, their impact, and the fix for each. Audits are also included quarterly for ongoing marketing clients. The audit covers tag firing, conversion accuracy, attribution settings, platform connections, and UTM consistency.

About LTL Creative: LTL Creative provides Calgary marketing analytics and reporting for ambitious local businesses, specializing in analytics audits, troubleshooting, and data accuracy verification, delivered through Google Partner-certified strategists for owners and marketing leaders requiring measurable, trusted results.

Ready to Be Our Next Success Story with marketing data you can actually trust? LTL Creative helps Calgary businesses diagnose and fix analytics issues that quietly distort decisions, backed by Google Partner, Meta-certified, and CXL-trained specialists.

Connect with LTL Creative today to discuss your analytics troubleshooting.

Disclaimer: Results vary by business, industry, and market conditions. Statistics, platform data, and pricing referenced reflect current industry benchmarks and are subject to change.