The 7 Most Common Tracking Mistakes at SMBs

Double-counted conversions, missing consent integration, missed goals. A guide to auditing your GA4 setup with a clear head.

The 7 Most Common Tracking Mistakes at SMBs

Bad data is worse than no data. When your tracking is broken, you don't just miss insights—you make confident decisions based on wrong numbers. We audit analytics setups for small and medium-sized businesses regularly, and these seven mistakes show up in nearly every one.

1. Double-Counting Conversions

This is the most common mistake and the most dangerous, because it makes your campaigns look better than they actually are.

How it happens: A conversion tag fires on the thank-you page, but the page reloads or the user navigates back and triggers it again. Or the same conversion is tracked by both Google Ads and GA4 with different attribution windows, and you add the numbers together.

How to fix it:

• Set conversion tags to fire only once per session or use transaction IDs to deduplicate
• Use a single source of truth for conversion counting—typically GA4—and import those conversions into Google Ads rather than tracking separately
• Audit your Tag Manager setup quarterly to catch duplicate triggers

2. Missing or Broken Consent Management

Since the GDPR enforcement wave and the rise of cookie consent platforms, many sites either block tracking entirely (losing all data) or implement consent incorrectly (collecting data illegally).

How it happens: The consent banner is misconfigured, so tags fire before consent is given. Or consent mode is set up but the fallback pings aren't enabled, meaning you lose 40-60% of your data with no modelling to fill the gaps.

How to fix it:

• Implement Google Consent Mode v2 properly—it allows Google to model conversions even when users decline cookies
• Test your consent setup with browser developer tools: clear cookies, visit your site, and verify that no tracking tags fire before you click "Accept"
• Review your CMP (Consent Management Platform) configuration at least twice a year

3. Wrong GA4 Configuration

GA4 is powerful but notoriously tricky to set up correctly. Most SMB installations we audit have significant configuration gaps.

Common problems:

• Enhanced measurement events are enabled but not useful (e.g., tracking every outbound link click as a "conversion")
• Data retention is set to 2 months instead of 14 months
• No custom dimensions or audiences are configured
• The Google Signals setting is on, which causes data thresholding (hiding your data) in reports

How to fix it:

• Set data retention to 14 months immediately
• Disable Google Signals unless you specifically need cross-device reporting
• Create custom events for the actions that actually matter to your business (not just pageviews)
• Set up GA4 audiences for remarketing from day one

4. No Server-Side Tracking

Browser-based tracking is increasingly unreliable. Ad blockers, ITP (Intelligent Tracking Prevention) in Safari, and privacy browsers all chip away at your data quality. If you rely solely on client-side tracking, you're likely missing 20-35% of your actual conversions.

How to fix it:

• Set up a server-side Google Tag Manager container. This runs on your own domain, which means tracking requests aren't blocked by ad blockers
• Implement the Meta Conversions API if you run Facebook/Instagram ads
• Use server-side tracking for critical conversion events while leaving less important events client-side to manage cost

5. Cross-Domain Tracking Issues

If your website sends users to a different domain for checkout, booking, or payment (common with Shopify, Calendly, Typeform, or payment gateways), the session breaks. GA4 treats the return visit as a new user from a "referral," and your conversion attribution is completely wrong.

How to fix it:

• Configure cross-domain tracking in GA4 by adding all relevant domains to your data stream settings
• Verify with the GA4 DebugView that the client ID persists across domains
• Add referral exclusions for any payment or booking domains so they don't overwrite your original traffic source

6. No Internal Traffic Filtering

Your own team visits your website dozens of times a day—testing, checking content, showing it to colleagues. Without filters, this internal traffic inflates pageviews, skews engagement metrics, and pollutes conversion data.

How to fix it:

• Define internal traffic in GA4 using IP addresses (Settings → Data Streams → Configure Tag Settings → Define Internal Traffic)
• Create a data filter to exclude internal traffic from reports
• If your team works remotely (no fixed IPs), use a custom dimension triggered by a URL parameter or browser extension

7. No Meaningful Conversion Goals

We see this constantly: GA4 is installed, but the only "conversion" is a pageview of the homepage. Or there are no conversions defined at all. Without meaningful goals, analytics is just a vanity dashboard.

How to fix it:

• Define 3-5 conversion events that represent real business value: form submissions, phone calls, demo requests, purchases, quote requests
• Assign monetary values to each conversion, even if they're estimates. This enables ROI reporting in Google Ads
• Set up micro-conversions too (e.g., scrolled to pricing, clicked "Call now," downloaded a PDF). These help you understand the path to conversion

The Compounding Effect of Bad Data

Here's the thing about tracking mistakes: they compound. Double-counted conversions lead to inflated ROAS, which leads to increased ad spend on underperforming campaigns, which leads to budget waste, which leads to the conclusion that "digital marketing doesn't work for us."

Meanwhile, the business down the road with clean data is optimising quietly, reallocating budget to what actually works, and pulling ahead.

A properly configured analytics setup isn't a nice-to-have. It's the foundation every marketing decision should rest on.

We fix these issues regularly as part of our analytics and tracking services. If you suspect your data might be off, reach out for a free quick audit—we'll tell you exactly what needs fixing.