GA4 Errors in Small Businesses: 8 Mistakes Skewing Your Data

Table of contents

• Why your GA4 numbers might be wrong without you knowing
• Mistake 1: the GA4 tag installed twice (plugin + GTM)
• Mistake 2: internal traffic left in, inflating your sessions
• Mistake 3: GDPR consent and Consent Mode V2 mishandled
• Mistake 4: events and conversions poorly named or missing
• Mistake 5: data retention left at the default 2 months
• Mistake 6: inconsistent UTMs breaking source attribution
• Mistake 7: spam, fake domains, and unfiltered traffic
• Mistake 8: reading GA4 like Universal Analytics (sessions vs. events)
• Keeping your GA4 accurate and monitored over time with Lysible
• Frequently asked questions

Why your GA4 numbers might be wrong without you knowing

A misconfigured Google Analytics 4 account never throws an error message. It keeps showing credible numbers that steer your decisions in the wrong direction. That's the real danger. You cut a profitable campaign. You double a budget on a phantom source. You never suspect the measurement was wrong to begin with.

When I open an account for an audit, I spot the main problem in under ten minutes. The pattern repeats itself: the property was set up in a hurry, often by a rushed developer or through a plugin, and nobody touched the settings afterward. GA4's flaws are configuration flaws, not bugs. The difference matters: a bug announces itself. A bad configuration doesn't.

The real cost is trust. According to the France Num 2024 barometer on the digital maturity of small businesses, a majority of small companies say they're poorly equipped to make use of their data. When the underlying data is wrong on top of that, the gap between decisions and reality becomes enormous.

How do I know if my GA4 data is reliable?

Three conditions need to be met: a single active tag, internal traffic excluded, and consent correctly passed through. On the flip side, an abnormally low bounce rate, sessions spiking for no reason on weekends, or zero conversions while you're actually selling: these are signals worth investigating.

The fastest test: open the real-time report, browse your own site, and check whether you show up more than once or with doubled pageviews. Thirty seconds, and you already know if the tag is clean.

The 3 symptoms of a poorly configured GA4

Three signals almost always give away a problem account, and you can spot them without any technical skill.

• A collapsed bounce rate (under 15%): the classic sign of pageviews being counted twice.
• Sessions climbing with no campaign behind them: internal traffic, bots, or referral spam inflating your volumes.
• Conversions stuck at zero while the business is running fine: events poorly named, not marked as conversions, or consent blocking collection.

You tick even one of these boxes? The rest of this article is for you. For a broader look at reading your reports, our guide on the GA4 reports that actually matter rounds out this diagnosis nicely.

Mistake 1: the GA4 tag installed twice (plugin + GTM)

Installing the tag twice is the number one mistake on small business accounts. It mechanically doubles your pageviews. It happens when the same measurement ID (the code starting with G-) gets placed in two spots: once through a plugin or the site theme, once through Google Tag Manager.

The symptom is clear. Your pageviews show double the real number. Your bounce rate collapses, because GA4 logs two interactions instead of one and assumes the visitor "engaged" with the page. Your session durations get thrown off too. Your whole report becomes wrong in the most misleading way possible: the numbers look flattering.

On a showcase site for an industrial small business audited last year, the account showed 40,000 monthly pageviews. After removing the duplicate tag, it dropped to 21,000. Half that "impressive" traffic never existed. The owner had justified a paid search budget on those inflated numbers.

How do I check if my tag is duplicated?

Open Google's Tag Assistant Chrome extension, or your browser's DevTools network tab, and load a page. If you see two requests firing to collect with the same G-XXXXXXX identifier, your tag is installed twice.

The fix is simple. Keep a single source of installation: if you use Google Tag Manager, disable the plugin or theme tag. If you don't use Tag Manager, leave the plugin as the only source. Never run both. Once the duplicate is removed, compare pageviews for 7 days before and after: the drop you see is the noise you were wrongly measuring.

Mistake 2: internal traffic left in, inflating your sessions

On a small site, your own team can account for 15 to 30% of sessions. That's enough to invalidate any marketing decision built on the data. Generic guides skip over this because they're written with large volumes in mind. For a small business, the effect is massive.

Let's run the numbers. A site with 3,000 monthly sessions, an 8-person team checking the site several times a day: each of them easily generates 40 to 80 sessions a month between testing, order checks, and simple curiosity. You quickly reach 500 internal sessions out of 3,000. Your conversion rates, your top pages, your user journeys: all of it is polluted by people who know the site by heart.

The symptom: admin or back-office pages showing up among your top pages, a mass of very short sessions, and a regular spike during office hours.

How much traffic does a small team account for?

For a site under 5,000 sessions a month, an active team of 5 to 10 people commonly accounts for 15 to 30% of measured traffic. The lower your traffic, the more overwhelming the relative weight of internal visits becomes.

The official fix is an internal data filter. You define a rule based on your IP address range in the data stream settings, then activate the filter. The exact steps are described in Google Analytics's help on excluding internal traffic. Crucial point: the filter defaults to "testing" mode. You have to switch it to "active" for it to actually work. I've seen dozens of accounts where the filter existed but had never been activated.

Mistake 3: GDPR consent and Consent Mode V2 mishandled

Without Consent Mode V2, you lose a significant share of data from European visitors who decline cookies, and you expose yourself to compliance risk. Google presents this setting as a technical box to tick. For a small business, it's a double issue: lost measurement and CNIL compliance.

Here's the tension nobody explains clearly. GDPR and CNIL doctrine require collecting consent before placing measurement cookies. If the visitor declines, GA4 by default stops collecting anything about them. And a meaningful share of visitors decline. Consent Mode V2 lets you recover modeled, anonymous, aggregated data even when the user has refused cookies. Without it, those visitors simply become invisible.

Since March 2024, Consent Mode V2 has been required to keep feeding Google Ads' advertising features and remarketing in the European Economic Area. A small business running ads without it installed loses both analytics data and campaign performance.

What happens if I haven't installed Consent Mode V2?

Two direct consequences. First, visitors who decline cookies vanish entirely from your stats: you underestimate your real traffic and your real conversions. Second, your remarketing audiences in Google Ads shrink, which hurts your retargeting.

The fix: connect your consent banner (Axeptio, Cookiebot, Tarteaucitron, or a certified equivalent) to Consent Mode through Google Tag Manager. Then check GA4's consent diagnostics to confirm the signals are coming through. The CNIL's reference page on cookies and trackers sets out the legal framework you need to follow. Don't wing this one: it's the only item on this list that carries a risk of penalties, not just skewed measurement.

Mistake 4: events and conversions poorly named or missing

An inconsistent naming convention makes your conversions unusable. It's often the reason "nothing shows up" in GA4. The whole model is built on events. If you name them haphazardly, you end up with Contact, contact_form, Formulaire, lead_ok all describing the same action, scattered across four different rows.

The problem gets worse with conversions. An event isn't automatically a conversion in GA4: you have to explicitly mark it as a "key event." Many small businesses track clicks on a purchase button without ever checking that box, then wonder why the conversions report is empty.

Why aren't my conversions showing up in GA4?

Your conversions aren't showing up for one of three reasons: the event doesn't exist (never set up), it exists but isn't marked as a key event, or consent is blocking it. Check in that order.

Here's the naming convention we apply on accounts we support, calibrated for a small business with no data analyst. It comes down to three rules: verb_object, in English, lowercase, no accents.

Lock in this standard before you create a single event. Once your events are clean, mark the most strategic ones as conversions. Our guide on GA4 conversion tracking for e-commerce covers setting up the purchase event in more detail.

Mistake 5: data retention left at the default 2 months

GA4 defaults to keeping detailed data for 2 months. The result: you can't compare year over year. It's a silent trap. Your standard reports stay accessible, but the moment you want to dig into past behavior in explorations, the data is gone.

Imagine wanting to compare your December traffic with the previous year's in a custom exploration. Impossible: GA4 has already erased anything older than 2 months. For a seasonal business (e-commerce, tourism, training), that's a dealbreaker. You lose any read on long-term trends.

The fix takes two clicks. Go to Administration, then Data Retention, and switch the setting from 2 months to 14 months. That's the maximum in the free version, and it's enough to compare one season with another. Do it today: this setting isn't retroactive, and data already erased won't come back. The details are in Google Analytics help on data retention.

Having opened dozens of GA4 accounts over the past five years, I can count on one hand the ones where this setting had ever been touched. Hands down, the best time-to-impact ratio on this whole list.

Lysible automatically checks this kind of setting on the accounts it connects to, so you don't have to remember to think about it once a year.

Mistake 6: inconsistent UTMs breaking source attribution

UTM parameters written without a consistent rule split a single campaign across multiple rows and make attribution unreadable. Newsletter, newsletter, NL_june, email: GA4 treats these as four different sources when it's really the same channel.

GA4 is case-sensitive. Facebook and facebook become two separate sources in your reports. Multiply that by the number of people creating links across your team, and your acquisition report turns into an unusable patchwork. You no longer know which channel is actually driving your sales.

Why aren't my UTM parameters working in GA4?

Your UTMs "don't work" most often because they're inconsistent (varying case, multiple values for the same channel) or misplaced. GA4 requires at least utm_source and utm_medium to correctly attribute a visit to its source.

The standard to enforce: all lowercase, no accents, no spaces (replaced with an underscore), and one reference sheet shared across the whole team.

One shared file that everyone copies their conventions from, and the problem disappears. To understand how GA4 then connects these sources to your sales, our article on the digital marketing attribution model explains the shift from last-click to data-driven.

Mistake 7: spam, fake domains, and unfiltered traffic

Referral spam and bots inflate a small site's reports with junk traffic. This noise is proportionally more visible when real traffic is low. On a large site, a few hundred bot visits get lost in the crowd. On a site with 2,000 sessions, they skew everything.

The classic symptom: oddly named sources in your acquisition report, often sketchy domains with 100% bounce rates and zero engagement. Sometimes also sudden spikes in unexplained "direct" traffic, or sessions coming from countries you don't sell in.

GA4 already filters some known bot traffic through the IAB list, but it doesn't catch everything. Three actions cut the noise effectively.

• Exclude unwanted referral domains in your data stream settings, to remove recurring junk sources.
• Set up your list of unwanted referral exclusions (your own subdomains, your payment provider) so Stripe or PayPal doesn't show up as the "source" of a sale.
• Regularly monitor your acquisition report to catch a new suspicious source before it pollutes your totals.

On a small e-commerce account audited recently, the displayed conversion rate was 0.4%. After cleaning out the junk traffic, it climbed back to 1.1%. Same business, same period, just the noise removed. A complete cleanup method is in our guide on web data analysis for faster decisions.

Mistake 8: reading GA4 like Universal Analytics (sessions vs. events)

The biggest interpretation mistake in small businesses is looking for Universal Analytics' old landmarks inside GA4. The model has changed. You have to unlearn sessions and start thinking in events. People say GA4 is complicated. Except GA4 isn't complicated: it's different, and it feels complicated because we're asking it the wrong questions.

Universal Analytics counted sessions and pageviews. GA4 counts events: every interaction (pageview, click, scroll, purchase) is an event. That shift changes everything from your very first look at a report. UA's famous "bounce rate" no longer means the same thing: GA4 thinks primarily in terms of engagement rate, meaning the share of sessions where the visitor actually did something.

Why is GA4 so hard to use?

GA4 feels hard because you're looking for metrics that no longer exist and an interface you used to know by heart. Once you accept the event-based model, the tool starts to make sense. The difficulty is a matter of habit, not real complexity.

Here's the translation table I give business owners so they stop hunting for the old landmarks.

Letting go of the old world frees up time and avoids misreadings. If you're just starting out, our getting started with GA4 for beginners guide walks through this logic step by step. And cross-referencing GA4 with Search Console to connect traffic and queries remains, according to Google Search Central's official documentation, the foundation of a healthy read on your visibility.

Keeping your GA4 accurate and monitored over time with Lysible

Fixing these 8 mistakes once is a good start. The real challenge is keeping the account clean over time: a plugin reactivates, a developer places a tag again, a campaign gets mistagged, and the noise creeps back in. That's exactly where Lysible offers a concrete answer for small businesses without in-house data skills. The platform centralizes your GA4, Search Console, and Google Ads data, flags measurement anomalies, and turns your numbers into clear decisions, without juggling ten different tools. You check the health of your tracking at a glance, instead of discovering a polluted account six months too late, right when you need to decide on your budget.