Data on Website Traffic: Where It Comes From, How Accurate It Is and What to Trust

data on website traffic

Traffic numbers come from very different places: your own analytics, Google Search Console, server logs, and third-party estimators that guess at other people’s sites. They rarely agree, and often for good reasons. This guide explains what each source measures, why Search Console clicks and Analytics sessions differ, how accurate competitor traffic estimates are according to published studies, and a practical way to build a single trustworthy view of your own traffic.

first-party vs third-party data
why numbers differ
accuracy studies
a simple reporting setup

// note: This guide summarises Google documentation and published studies. I have not run my own accuracy test of traffic estimators. Where I cite a study, I name its author and sample size, and say when I could only read a summary.

traffic-sources.txtsummary

source	measures	trust level
Search Console	clicks from Google	high for Google search
GA4	sessions on your site	high if tagged well
Server logs	all requests incl. bots	high but noisy
Third-party tools	modelled estimates	directional only

// see the sections below for evidence.

key takeaways

The Short Version Before the Deep Dive

// 1

First-party beats estimates

For your own site, Search Console and a properly installed analytics tool are the ground truth. Estimators are for sites you cannot measure.

// 2

Clicks are not sessions

Search Console counts clicks on Google results. Analytics counts sessions on your site. They answer different questions and should not match exactly.

// 3

Estimates can be far off

Ahrefs’ own study of 1,635 sites reported a median deviation of about 50 percent from Search Console. Treat competitor traffic as a rough guide.

sources

The Four Kinds of Traffic Data

Know which one you are looking at before you compare numbers.

Source	How it collects data	Strength	Blind spots
Google Search Console	Google records impressions and clicks in Search, Discover and News	Direct data on how people reach you from Google; queries and positions	Google only; no on-site behaviour; sampled link and query detail
Google Analytics 4	A script or tag on your pages records visits, events and conversions	All channels; behaviour and conversions after the click	Blocked by ad blockers, consent choices, broken tags; non-HTML files not counted
Server or CDN logs	Every request to your server, including bots	Independent of browser scripts; shows crawler activity	Includes non-human requests; needs interpretation
Third-party estimators	Models built from keyword rankings, clickstream panels and partner data	The only view of competitor sites	Modelled, can be far from reality, weak for small sites

SOURCE 1

What Search Console Actually Counts

Google’s documentation defines the Performance report metrics precisely (Google Search Console Help):

Impressions: how often a link to your site appeared in a results page, even if it was not scrolled into view. Items inside carousels or expandable sections generally need to be scrolled into view or expanded.
Clicks: clicks on a Google result that take someone to your site. Query refinements do not count, and returning to the same link and clicking again counts once.
Average position: the topmost position a link held, averaged across queries, where 1 is the top spot. Ads do not occupy a position.
CTR: clicks divided by impressions.

Data is attributed to the canonical URL Google picks, so mobile and desktop versions are combined. That matters when you compare it with an analytics tool that records the URL that actually loaded.

// when to use GSC

Best questions it answers

Which queries bring people to my pages?
Did a page lose impressions or position?
Is my click-through rate low for a good position?
Which pages Google is actually showing?

SOURCE 2

What Google Analytics 4 Measures and Forgets

GA4 records what happens after a visitor loads a page that carries its tag: sessions, engagement, events and conversions. It covers every channel, not just Google search.

One setting to know about is data retention. Google says GA4 event-level data can be kept for 2 months or 14 months on the standard version, with longer periods only in Analytics 360. Age, gender and interest data are always kept for two months. Retention affects explorations and funnel reports, not the standard aggregated reports (Google Analytics Help).

Practical consequence: if your property is still on the default 2 months, a year-on-year exploration will silently hide older data. Check the setting under Admin, then Data collection and modification, then Data retention. I would set 14 months.

ga4-health-check.txt

$ checks before trusting GA4
[ ] tag fires on every page
[ ] own IP / staff traffic excluded
[ ] data retention set to 14 months
[ ] key events defined
[ ] Search Console linked
[ ] referral exclusions configured

why numbers differ

Why Search Console Clicks Never Match GA4 Sessions

Different tools measure different moments. The gap is normal.

// what each measures

Clicks vs sessions

A click is counted when someone clicks a Google result. A session is counted when a page loads and the analytics tag runs, adjusted for timeouts. One person can create several sessions from one click, or leave before the tag fires.

// when GSC is higher

Clicks without sessions

Visitors leave before the page and tag load
Ad blockers or cookie consent stop tracking
Non-HTML files such as PDFs or images get clicks but run no tag
Bot clicks that never execute scripts

// when GA4 is higher

Sessions without clicks

Organic sessions from other search engines are included unless you filter to Google
Returning visitors from bookmarks may be attributed to the last non-direct source
Misconfigured cross-domain tracking can duplicate sessions

// other mismatches

Time and attribution

Search Console reports on Pacific time; GA4 uses your property time zone
Landing page URL can differ from the canonical URL Google credits
Google notes some processing differences remove duplicates and robot visits

// sources: Search Engine Journal, Bounteous, Swydo. The time zone and attribution points come from a vendor article, so verify them in your own account.

The takeaway: compare trends, not totals. If both lines rise and fall together, your tracking is probably healthy. If they diverge sharply after a site change, investigate the tag, consent banner or redirects.

third-party data

How Accurate Are Competitor Traffic Estimates?

Short answer: directionally useful, numerically unreliable, especially for smaller sites.

Tools such as Ahrefs, Semrush and Similarweb never see another site’s server logs. They infer traffic from keyword rankings and search volumes, from clickstream panels, or from partner data. Each method has gaps.

Evidence from published tests:

Ahrefs study: across 1,635 random sites, Ahrefs compared its US organic estimate with each site’s Search Console traffic and reported a median deviation of 49.52 percent. The same comparison put Semrush at 68.36 percent, according to a summary on ReWork (Ahrefs). That is a vendor testing its own product and a competitor, so read it with care.
SparkToro study: Rand Fishkin compared Semrush, Datos, Similarweb and Ahrefs against Google Analytics to see which correlated best, as described in coverage of the study. Critics argued it mixed traffic types, because Ahrefs estimates organic search only.
Smaller independent tests: a 787-site comparison cited by CustomerThink found Semrush and Similarweb each off by roughly 45 to 61 percent, and a Practical Ecommerce writer reported Ahrefs and Semrush estimates of his own site were more than 50 percent below actual.

Across all of these, small sites are the weak spot, because there is less data to model from. One 2026 tool-review blog suggests treating Similarweb as directional below about 500,000 monthly visits, which fits the pattern but is an opinion, not a measured threshold.

// how to use estimates safely

Rules of thumb

Compare ratios (you vs competitor) rather than absolute visits
Use trends, not single months
Check the same competitor in two tools
Prefer organic estimates for SEO decisions
Never put an estimate in a client report as fact
Validate against Search Console for your own site to learn the tool’s bias

QUALITY

Bots, Consent and Other Distortions

Even first-party data can mislead:

Bot traffic: many bots never run analytics scripts, so they show in logs and sometimes in Search Console but not in GA4. Others do run scripts and inflate sessions with near-zero engagement.
Consent banners: if visitors decline tracking, GA4 sees fewer of them while Search Console still counts their clicks.
Ad blockers: block the script entirely for many visitors.
Referral spam and internal traffic: staff visits and test pages inflate small sites.
Redesigns: a theme change can drop the tag and look like a traffic collapse.

If you see a sudden drop in one tool only, first check tracking before assuming a ranking problem. If the drop is real in Search Console, see my indexing guide and Penguin guide for causes to rule out.

// triage order

Traffic fell, now what?

Check the tag is firing
Compare with Search Console
Check date alignment and seasonality
Check index coverage and manual actions
Look at the pages and queries that dropped

setup

A Simple Reporting Setup for a Small Site

Verify the site in Search Console. Use a domain property so every protocol and subdomain is covered.
Install GA4 and link it to Search Console. That lets you see queries and landing pages together.
Set retention to 14 months and mark real conversions, such as form submissions or calls, as key events.
Filter your own traffic by IP or with a developer-traffic filter.
Pick three numbers for the monthly report: Search Console clicks, GA4 organic sessions, and conversions from organic. Add impressions and average position for context.
Annotate changes such as redesigns, migrations and campaigns so you can explain spikes later.
Use an estimator only for competitors, and record the date and tool alongside any figure.

If you want a proper written report format, see my SEO report guide. For tool options, see the best SEO tools roundup.

what to track

Metrics That Matter More Than Raw Traffic

// quality

Engaged sessions and conversions

Traffic without leads or sales is a vanity number. Track conversions from organic traffic.

// visibility

Impressions and queries

Impressions show if you are being seen even when clicks lag. New queries signal topical growth.

// efficiency

CTR by page

Pages with good positions but low click-through often need a better title and description.

// content

Landing page mix

See which pages bring visitors, and which of those convert.

// local

Calls and directions

For local businesses, Business Profile actions often matter more than site sessions.

// trend

Year over year

Compare the same months across years to remove seasonality. This is where 14-month retention helps.

mistakes

Common Traffic Data Mistakes

Avoid

Reporting competitor estimates as facts
Comparing clicks in one tool with sessions in another and calling it an error
Judging SEO by total visits instead of conversions
Ignoring seasonality
Changing the tracking setup without annotating it

Do instead

State which tool and metric each number comes from
Compare like with like over time
Track leads and revenue from organic
Use year-over-year views
Keep a change log

// how this guide was made

What I Checked and What I Didn’t

I read Google’s Search Console documentation on performance data, Google Analytics documentation on retention, and published articles on the Search Console versus Analytics gap and on traffic estimator accuracy. I could not open the full SparkToro or Authority Hacker studies, so I rely on summaries of them. I did not run my own accuracy test. Percentages are quoted as the publishers state them, and vendor-run tests are flagged as such.

faq

Website Traffic Data FAQ

Where can I get accurate data on my own website traffic?

Use Google Search Console for visits from Google search and Google Analytics 4 for sessions and conversions on your site. Server logs add an independent view including bots.

Why do Search Console and Google Analytics show different traffic?

They measure different things: clicks on Google results versus sessions on your site. Ad blockers, consent, bots, non-HTML files, time zones and attribution all create gaps.

How long does GA4 keep my data?

Event-level data is kept for 2 or 14 months on standard properties. This affects explorations and funnels, not standard reports.

How accurate are website traffic estimators?

Not very. Ahrefs reported a median deviation of about 50 percent from Search Console in its own 1,635-site study, and other tests show large gaps, especially for small sites.

Can I see a competitor’s exact traffic?

No. Only the site owner has first-party data. Tools provide estimates, so use them for comparison and trends.

What is a good source for industry traffic benchmarks?

Treat benchmarks cautiously. Compare against your own history and your closest competitors rather than broad averages.

Is Search Console data complete?

Its performance data is Google’s own count of impressions and clicks in Google surfaces, but it covers Google only and its link reports are samples.

How do I know if my traffic is mostly bots?

Look for sessions with zero engagement, odd geographies or user agents, and spikes unrelated to anything you did. Server logs make bot activity easier to see.

What should I report to my boss or client?

Search Console clicks, GA4 organic sessions and conversions from organic, each labelled with its source and compared month on month and year on year.

sources

Sources Used on This Page

// keep reading

More on Measurement

See the SEO report guide, SEO management and the best SEO tools roundup.