Why Don't My Analytics Tools Show the Same Numbers? (Fix Guide)

If you have ever installed two analytics tools and watched the numbers diverge by 30%, you already know the surprise. Open GA4 and Sleek side by side on the same site and the visitor counts will not match. Open Sleek and your server logs and they will not match. Open GA4 and Google Search Console and those will not match either.

This is not a bug in any specific tool. Different products measure with different methodologies, filter different traffic, attribute differently, and sample differently. The question is not "which tool has the right number" — the question is "which tool is closest to the question I actually want to answer?"

This is by far the largest source of discrepancy on technical sites. Ad blockers (uBlock Origin, Brave's built-in blocker, Firefox's Enhanced Tracking Protection, NextDNS, Pi-hole) block Google Analytics as a category. They typically do not block privacy-friendly analytics like Sleek, Plausible, or Fathom because those scripts are not part of the ad-tech filter lists.

Plausible measured this in 2021: 58% of Hacker News visitors blocked GA. On a SaaS marketing site with developer audience, GA4 may be missing 30–50% of real traffic.

On a non-technical audience (lifestyle blog, ecommerce mass-market), the blocking rate is lower — 5–15% — but still meaningful.

In the EU, GDPR + ePrivacy require opt-in consent for non-essential cookies. GA4 sets cookies, so EU traffic that declines consent is either not tracked or tracked with reduced accuracy via Google Consent Mode v2 modeled data.

Real-world consent rates in the EU vary wildly: a clean banner with prominent Accept can hit 60–80%, but legally compliant banners with equal-prominence Reject typically see 30–50% acceptance. That means GA4 measures only 30–50% of EU traffic at full fidelity; the rest is modeled or missing.

Privacy-friendly analytics do not face this issue because they do not need consent. Every EU visitor is counted at full fidelity.

Bots are a significant share of internet traffic — search engine crawlers, monitoring bots, scrapers, malicious bots. Different tools filter them differently.

GA4 does some bot filtering against the IAB/ABC International Spiders & Bots List, but it is conservative — it removes obvious crawlers but leaves a lot of automated traffic in the data. This is why GA4 sometimes shows mysterious traffic spikes from one country or a strange page distribution.

Privacy-friendly analytics like Sleek and Plausible filter bots aggressively — they reject anything matching known bot patterns, anything without typical browser fingerprints, and anything that looks automated. The result: cleaner data but slightly lower counts.

When GA4 shows higher numbers than a privacy-friendly tool, it is often because GA4 is including bots the other tool filtered out. Server logs will reveal which is closer to truth.

When a visitor arrives via paid ad on Monday and converts on Friday, who gets credit? GA4 defaults to a 7-day click and 1-day view attribution; Facebook Ads defaults to 7-day click and 1-day view; Google Ads can extend to 90 days; Plausible/Sleek attribute to the most recent touch in the same session.

For a single conversion, three tools can show three different sources. None are wrong — they are answering different questions. GA4: "what was the source within 7 days of conversion?" Sleek: "what was the immediate referrer?" Facebook: "did Facebook contribute within 7 days?" The numbers will not reconcile because the questions do not.

GA4 free tier samples your data when reports exceed certain thresholds (typically 10M+ events in a custom date range). Sampled reports can differ from the underlying data by several percent.

Privacy-friendly analytics typically do not sample because their data volumes per customer are smaller and their architecture handles full-fidelity reporting. But on very large sites, even Sleek may aggregate at the day level rather than the second level for performance.

If you are seeing different numbers on different date ranges in the same tool, sampling is a likely cause. Check the report settings.

A "session" or "visit" is a fuzzy concept. GA4 starts a new session when the user changes source/medium or after 30 minutes of inactivity. Plausible counts a unique visit as the same person in the same day. Sleek treats a session as activity within 30 minutes of inactivity, regardless of source.

For one user who arrives, leaves, and comes back via a different source the same day, GA4 might count 2 sessions; Plausible 1 visit; Sleek 1 or 2 depending on the gap. None of these are wrong — they are answering different questions about engagement.

Two privacy-friendly tools running on the same site with similar bot filters should land within 3–5% of each other. If Sleek and Plausible disagree by 15%, something is wrong — usually a tracking script not loading on every page, or one tool installed on a subdomain the other is not.

Two GA4 properties pointing at the same site should match within 1%. If they don't, you have duplicate tags firing or an attribution config difference.

GA4 and a privacy-friendly tool will not match. That is normal. Stop trying to make them match — pick one as the source of truth and accept the others as additional perspectives.