The Only Website Metrics That Matter (And How to Track Them Free)

Where your traffic really comes from

Channel breakdowns matter less than they used to, because privacy changes and app browsers hide a lot, but they still give you directional insight.

You just have to read them with a bit of skepticism.

• Organic search: visitors who found you on Google, Bing, or other engines.
• Direct: people who typed your URL, used bookmarks, or arrived without referrer data, which also includes a chunk of “dark” traffic from apps and email.
• Paid: traffic from your ads when tagged correctly with UTM parameters.
• Referral: links from other websites, forums, and some apps.
• Organic social: visits from platforms like LinkedIn or Facebook, when their referrers are not stripped.

AI tools sit in a weird place.

Some, like Perplexity, show up as normal referrers; others act more like a search engine with zero clicks, where you get visibility but no visit in your analytics account.

Time, location, and device patterns

The classic filters still help you understand behavior, even if some slices of traffic are hidden by consent banners or ad blockers.

I would still check three simple breakdowns at least once a month.

• Country and city: where your best converting sessions come from, not just where most visits come from.
• Device category: desktop, mobile, tablet; watch conversion rate by device, not just volume.
• Time of day / day of week: when engaged sessions and conversions peak, which helps with email timing and ad scheduling.

When mobile users bring most of your conversions but see lower engagement rates than desktop, that is usually a design or performance problem, not “mobile users are flaky”.

For example, one SaaS site I worked with had 65 percent of users on mobile but almost all trials started from desktop.

Fixing mobile load time and form friction lifted mobile trial starts by over 40 percent in a few months, with no extra traffic at all.

Mind the data gaps from privacy and consent

This is the part many reports quietly ignore.

Your real traffic is higher than what almost any analytics tool shows.

Ad blockers, do‑not‑track signals, and consent refusals all remove a slice of users from your dashboards.

That slice is different per country and per audience, so chasing perfect accuracy is a waste of time.

Instead, zoom out a bit.

Look at trends and percentage changes across months, not single days, and compare channels against each other more than against some imagined “real” number.

Engagement metrics that actually matter now

Engagement is where most people still think in old terms like “bounce rate” and “time on site”, which do not quite match how GA4 thinks.

If you are not careful, you can read a GA4 report using a Universal Analytics mental model and draw the wrong conclusion.

Engaged sessions and engagement rate

In GA4, engagement rate is the default way to see whether visits are real visits or just drive‑bys.

An engaged session in GA4 is counted when any of these happen:

• The session lasts 10 seconds or longer.
• The user views at least 2 screens or pages.
• The session triggers a conversion event.

Engagement rate is simply the share of sessions that meet those rules.

If engagement rate is 60 percent, it means 60 out of 100 sessions were engaged by that definition.

GA4 also offers a new “bounce rate”, but it is just 100 percent minus engagement rate.

I would not lead with it; the engaged view is just more helpful mentally.

What a good engagement rate looks like

Benchmarks are always rough, so use these as a starting point, not a final grade.

If your engagement rate drops sharply after a redesign or a big content change, that is a signal worth taking seriously.

If it moves a couple of points week to week, that is usually just noise.

Average engagement time vs session duration

Universal Analytics tracked “session duration” and “time on page”, but those metrics were always a bit misleading.

GA4 switched to “average engagement time”, which only counts when the page is in the foreground and there is some activity.

Average engagement time per session is more realistic, but you still should not worship it.

If a page answers a question in 20 seconds and sends a user to your checkout, that short time is a win, not a problem.

Use engagement time along with two simple checks:

• Is the conversion rate for this page or funnel step strong enough?
• Do scroll depth and key events show people reaching the content that matters?

Tools like Microsoft Clarity, with scroll heatmaps and click maps, can fill in the story that raw time cannot tell alone.

I like to pair GA4 numbers with Clarity once a month on my top 10 pages, just to keep a real‑world sense of behavior.

Views per user and navigation quality

Views per user (or per session) used to be praised as a simple sign of “stickiness”.

That can still be true, but it depends on what your site is trying to do.

Examples:

• If you run a content site with ad revenue, more views per session can be good, up to the point where users feel trapped.
• If you run ecommerce, a user who finds what they want in 3 views is better than one who needs 10 clicks to find the same product.
• For B2B lead gen, you want the right mix: a few product pages, a comparison page, maybe a case study, then a contact or signup.

Instead of chasing a single magic number for pages per session, outline what an ideal pre‑conversion path looks like for your business.

Then check how many sessions follow some version of that pattern, and work on making that path easier and more obvious.

Metrics you should stop obsessing over

If a metric does not move conversions, revenue, or clear leading indicators like engaged sessions, it is a vanity metric.

Here are a few I would keep in the background:

• Raw pageviews: fine for seeing popularity, but not a success metric by itself.
• Classic bounce rate: use engagement rate instead; it is a better default mental model.
• Average session duration in old reports: GA4 engagement time is more meaningful.
• Social reactions that never bring people to your site or your app.

I know it feels good to see giant numbers, but those do not pay salaries or fund growth.

Your reports will get cleaner and your decisions sharper once you drop half the metrics you have been trained to care about.

SEO, AI visibility, and the search metrics that count

Search is still the biggest long‑term driver of high intent visitors for most sites, but the way search looks now is pretty different from a few years ago.

Classic “rank tracking” alone is not enough, because two results in the same position can perform very differently in practice.

Impressions, clicks, CTR, and the new SERP reality

In Google Search Console you get three core metrics that matter more than any third‑party ranking tool.

If you only watch one SEO report, make it the performance report here.

• Impressions: how often your page appeared for a query.
• Clicks: how often people chose your result.
• Click‑through rate (CTR): clicks divided by impressions.

Average position is still visible, but take it with caution.

A position of 3 at the bottom of a huge AI overview, People Also Ask boxes, and a video carousel might get fewer clicks than a position 5 in a cleaner layout.

So I would track it like this:

• Watch impressions to see if Google trusts you enough to show you for the right topics.
• Watch CTR per page and per query to see if your title and description are winning the click.
• Look at clicks and conversions to decide where to update or create content.

AI Overviews, zero‑click answers, and brand visibility

AI Overviews and similar features can answer a user without a click, yet your content can still be used inside those answers.

That creates a bit of a paradox: you can influence the answer but not always see a visit.

There is no single perfect report for this, but you can piece together a useful picture.

• Watch impressions for key queries in Search Console; if impressions rise but clicks and CTR fall while conversions from that topic stay stable, some of the demand may be satisfied inside AI Overviews.
• Look for branded search impression growth: more branded impressions combined with stable or slightly lower generic clicks can still be a net positive for awareness.
• Use third‑party tools that track whether your pages are cited in AI answer panels or “read more” expansions, and log those for your top topics.

In practice, I care less about whether AI Overviews “steal” clicks and more about whether my brand keeps showing up where buyers start their research.

If your non‑branded impressions are flat for important terms while competitors surge, that is a real risk.

In that case, refreshing content depth, adding clear entities (brands, products, people, locations), and tightening factual accuracy can help you be a better source for both classic and AI‑generated results.

Tracking direct AI referrals

Some AI tools do send measurable traffic.

For example, Perplexity, certain AI browsers, and some ChatGPT integrations can show up in your referral reports.

To catch these, do two things:

• In GA4, go to Reports → Acquisition → Traffic acquisition, then break down by session source / medium and look for new or unusual referrers that look like AI tools or AI browsers.
• Create a custom channel group that labels those sources as “AI referrals” so you can track them over time instead of losing them in “referral” noise.

The numbers might be small now, but they will probably grow.

The real value is less about raw traffic and more about which topics tend to be cited and linked from those tools.

Backlinks and referring domains that matter in 2026

Link building is still a big part of SEO, but what counts as a good link has tightened after multiple spam updates.

I think the easiest mental model is simple: would a human, with no SEO awareness, see this link and think “this is helpful and relevant”?

Focus on:

• Links from sites that actually send some traffic.
• Links from pages closely related to your topic or niche.
• A healthy mix of domains, not hundreds from the same small cluster of sites.

Use tools like Ahrefs, Semrush, or even free reports from smaller platforms to watch new referring domains and see which content attracts the best links.

Then build more along those lines, instead of chasing random directory links or low quality guest posts.

Search appearance and schema

Search Console has a “Search appearance” section that many site owners ignore.

It shows how your pages appear in search: rich results, video results, and other formats.

Some rich result types have come and gone, but structured data is still worth your time for core types:

• Product markup for ecommerce, with price and availability.
• Article and FAQ markup where it still makes sense.
• Organization and person markup so your brand and authors look clear to search engines.

Track impressions and clicks by search appearance in GSC to see if your structured content is actually winning more visibility.

If a type of rich result vanishes or stops performing, do not cling to it; redirect that effort into deeper, clearer content on important pages.

Business outcome metrics: from clicks to revenue

This is the part most dashboards gloss over, but it is where smart reporting lives.

You do not run a business on traffic, you run it on revenue, pipeline, or product usage, and your website either supports that or it does not.

Conversions, conversion rate, and value per conversion

Start with clear conversion definitions for your site.

If you are vague about what counts, your metrics will be vague too.

• Ecommerce: purchases, maybe add‑to‑cart or start‑checkout as micro conversions.
• B2B lead gen: form submissions, demo requests, qualified inbound calls.
• SaaS: free trial starts, demo bookings, plan upgrades.
• Content / community: email signups, account creations, or key engagement milestones.

In GA4, you create events for these actions and mark the key ones as conversions under Admin → Events.

For ecommerce, you can pass actual revenue for each transaction; for lead gen, you can attach a simple average value per lead or per signup.

Once you have that, two numbers do a lot of heavy lifting:

• Conversion rate = conversions / sessions (or / users, depending on how you like to frame it).
• Value per session = revenue (or lead value) / sessions.

Value per session is underrated.

I think of it as “how much is one more visit worth to you, on average?” and it helps you sanity‑check ad spend and content investment.

CAC, lead quality, and tying into your CRM

Marketing costs need a match on the revenue side, or you are just guessing.

To tie things together, you usually need at least a lightweight handoff between analytics and your CRM or sales tool.

Here is a simple flow that works for many small and mid‑size teams:

• Track form submissions or signups as conversions in GA4 with a hidden field capturing basic source / medium / campaign info.
• Send that data into your CRM (HubSpot, Pipedrive, Salesforce, etc.) with the lead record.
• After some time, look at which channels actually create opportunities and closed deals, not just leads.

This lets you calculate:

• Customer acquisition cost (CAC) from web: total web marketing and ad spend divided by new customers from web in that period.
• Lead‑to‑sale rate by channel: sales / leads for organic, paid, referrals, and so on.

A channel that brings half as many leads but three times the close rate is often more valuable than it first appears in GA4.

If you only judge by on‑site conversion rate, you might kill the channel that sends your best customers.

Pulling in even basic CRM feedback keeps you from that mistake.

Cohorts, retention, and product‑influenced metrics

For SaaS, membership, or subscription products, the story does not end at signup.

Churn, expansion, and long‑term engagement matter at least as much as acquisition.

You can start simple:

• Track which acquisition channels bring users who activate in the product (hit a “success” action like completing onboarding, using a key feature, or inviting a teammate).
• Look at trial‑to‑paid conversion rate by landing page, not just by channel; your product pages and pricing layout influence this a lot.
• Watch help docs and onboarding content: pages that reduce support tickets or lower churn are worth more than their traffic suggests.

This is where many teams under‑report their wins.

A documentation revamp that reduces churn is as powerful as a new acquisition channel, but it rarely shows up in a simple traffic dashboard.

Example: turning visits into a simple revenue picture

Let me make this concrete with a rough example.

Say your B2B site had last month:

• 10,000 sessions.
• 2.5 percent form submission rate, so 250 leads.
• 20 percent of those leads became opportunities (50), and 20 closed (10 new customers).
• Average customer value in the first year: 4,000.

Your numbers now look like this:

• Revenue from web this month: 10 × 4,000 = 40,000.
• Revenue per session: 40,000 / 10,000 = 4.00.

If your total monthly web marketing cost was 12,000, your CAC from web is 12,000 / 10 = 1,200.

You can now ask a real question: “Can we reduce CAC or raise revenue per session through better traffic, better messaging, or a smoother funnel?” and design experiments around that, instead of chasing random traffic growth.

Technical health, Core Web Vitals, and access

Technical metrics are not sexy, but they often explain stubborn conversion problems and slow SEO growth.

Three things tend to matter most today: Core Web Vitals, indexing, and accessibility basics.

Core Web Vitals: LCP, INP, CLS

Core Web Vitals are Google’s preferred way of measuring load speed, responsiveness, and visual stability for real users.

You can see them in PageSpeed Insights and in the Experience section of Search Console.

• Largest Contentful Paint (LCP): how long it takes the main content to appear; aim for under 2.5 seconds for most users.
• Interaction to Next Paint (INP): how quickly the page responds when users interact; here you want most interactions under 200 ms.
• Cumulative Layout Shift (CLS): how much stuff jumps around while loading; lower is better, keep it under 0.1.

Instead of obsessing over tiny score changes, watch how many URLs fall into “good”, “needs improvement”, or “poor”.

Then focus on templates: product pages, blog posts, category pages, and so on, instead of every single URL.

Mobile performance and mobile‑first reality

Mobile‑first indexing is fully rolled out, which means Google mainly judges you by the mobile version of your pages.

Your users often do the same without saying it.

Check this in two quick ways:

• GA4: compare engagement rate and conversion rate on mobile vs desktop; big gaps should push you to test design and speed on mobile first.
• PageSpeed Insights: run tests with the mobile option, and watch Core Web Vitals specifically for mobile visitors.

Fixing mobile page weight, image sizes, and layout shifts can have a bigger effect on conversions than most copy tweaks.

On a few ecommerce stores, simple image compression plus removing heavy third‑party scripts lifted revenue more than new ad campaigns did.

Indexing and crawl issues

You can write the best content in your niche and still get no traffic if search engines cannot crawl or index it cleanly.

This is one of those boring tasks that saves you months of confusion later.

In Search Console, look at:

• Pages → Indexed: which pages are in the index and which are excluded.
• Reasons like “Crawled – currently not indexed” or “Discovered – currently not indexed”, which can hint at thin content, duplication, or crawl budget issues.
• Coverage patterns, such as many important URLs blocked by robots.txt or tagged with noindex by mistake.

Complement that with a crawl from a tool like Screaming Frog, and with basic server log checks or uptime monitoring so you catch outages and bot spikes that normal analytics might smooth over.

Accessibility as a metric

Accessibility is not only a legal or ethical topic, it is a practical UX metric.

If people with different abilities cannot use your site, your conversion stats quietly suffer.

You do not need to become a full accessibility expert, but you can track:

• Basic Lighthouse accessibility scores for your main templates.
• Alt text coverage for key images, especially on product and feature pages.
• Keyboard navigation on core flows like checkout and signup.

A small improvement here often overlaps with better mobile usability and better Core Web Vitals anyway.

So you are not really doing “extra” work, you are just fixing real friction your metrics did not show clearly before.

Tracking the right metrics free: a practical stack

You do not need expensive tools to track the metrics that matter.

A lean free stack, used consistently, beats a flashy paid one that nobody logs into.

Minimum viable analytics setup

If I had to start from scratch on a small site, I would set up just three tools.

If privacy laws in your region are strict, you can pair or replace GA4 with a self‑hosted tool like Matomo or a simpler option like Plausible, but the idea stays the same.

One tool for behavior and conversions, one for search, one for visual behavior diagnostics.

How to track core metrics in GA4, step by step

You do not need every GA4 feature to track the metrics we have talked about.

Here is a bare‑bones setup that covers most sites.

1. Install GA4 using gtag.js or Google Tag Manager on every page.
2. Set up key events:
   • For forms, fire an event like “form_submit” on successful submission.
   • For ecommerce, enable GA4 ecommerce events through your platform integration where possible.
   • For SaaS or apps, send events like “trial_started” or “plan_upgraded” from your backend or tag manager.
3. Mark conversions in Admin → Events: toggle on “Mark as conversion” for your main events.
4. Create basic reports: under Reports → Engagement → Events and Conversions, and under Reports → Acquisition for traffic by channel.
5. Build a simple exploration for qualified sessions, using engaged sessions plus your country and channel filters.

You can then set up email summaries or saved comparisons for your key segments so you do not need to dig from scratch every time.

Three simple monthly dashboards you can build

To keep reporting tight, I like to match the dashboard to the business model.

Here are three quick recipes.

Ecommerce

• Users and sessions by channel.
• Engagement rate by channel.
• Revenue, conversion rate, and average order value by channel.
• Top 10 products by revenue and by conversion rate.

B2B lead gen

• Users and engaged sessions by channel.
• Form submissions and conversion rate by landing page.
• Lead quality sample from CRM by channel (opportunities and wins).
• Top 10 content pages that helped before conversion (assisted pages).

SaaS / subscription

• Users and sessions by channel.
• Trial starts or signups by channel and landing page.
• Trial‑to‑paid rate by acquisition source, from your product or CRM data.
• Usage of key product features and visits to help docs or onboarding guides.

Notice how each list depends on very few metrics.

You can always add more later, but starting lean keeps you honest.

Competitor comparison and reading patterns over time

Your metrics only make full sense in context: against your past performance and against realistic competitors.

Otherwise you either feel behind for no reason or think you are winning when the whole market is shrinking.

Using third‑party estimates without fooling yourself

Tools like Similarweb, Semrush, and Ahrefs can estimate competitor traffic and top pages.

Those numbers are not exact, and that is fine as long as you treat them as trends, not as gospel.

Look for patterns like:

• Which channels each competitor leans on: organic, paid, social, referrals.
• Topics they cover more deeply than you do.
• Countries where they seem stronger than you.

If they beat you on content volume, your response might be better depth, better structure, or more focused topical clusters rather than trying to match every single article.

If they win on branded search, you may need more awareness campaigns, partnerships, or PR, not just SEO tweaks.

Spikes, dips, and long‑term trends

Inside your own data, the real insight often comes from changes, not snapshots.

Here are a few patterns worth watching.

• Sharp spikes in traffic with flat conversions: often caused by viral social posts, unqualified mentions, or bot traffic; interesting, but not always good.
• Slow, steady growth in organic impressions and engaged sessions: usually a healthier sign that content and technical work are paying off.
• Channel‑specific drops: a sudden fall in organic traffic paired with lost impressions may point to an algorithm change or stronger competitors; a fall in direct might hint at brand interest fading.
• Seasonal swings: retail, tax, travel, and education niches in particular have clear peaks; map your last couple of years to avoid misreading normal seasonality as crisis.

When something shifts, ask “what changed right before this line moved?” and check content updates, technical changes, and campaign launches before blaming the algorithm.

Over a year or two, your goal is simple: more qualified sessions, higher or stable conversion rates, and better value per session.

Everything in your reporting should help you see whether that story is moving in the right direction.

Putting all of this into a simple weekly habit

You do not need to live inside analytics to run a strong website, but you do need a short, consistent rhythm.

Think of it as a quick health check rather than a full exam every time.

A realistic review cadence

Here is a pattern that tends to work for most teams.

• Weekly, 15-30 minutes:
  • Check users, engaged sessions, and conversions by channel.
  • Glance at any alert‑worthy drops in Search Console coverage or impressions.
  • Review one or two top pages in Clarity to spot friction.
• Monthly, 60-90 minutes:
  • Update your simple dashboard for leadership or clients.
  • Compare this month against the previous month and the same month last year.
  • Pick 2-3 specific experiments based on what the metrics suggest.
• Quarterly:
  • Review Core Web Vitals, indexing, and key templates.
  • Revisit conversion tracking and CRM alignment for lead quality.
  • Do a light competitor review to see how far the goalposts have moved.

This rhythm keeps you close enough to reality without letting numbers dominate your week.

If a metric change would not affect what you do next week or next month, it probably does not deserve a daily check.

Where to focus next

If you feel a bit overloaded right now, pick just three moves to start with.

• Clean up your GA4 events and conversions so you can trust your core numbers.
• Connect Search Console and look at your top queries and pages by impressions and clicks.
• Run a Core Web Vitals and mobile usability check on your most important page types.

Once those are in place, everything else in this guide gets easier, because your metrics start to reflect how your website really supports your business.

From there, you can decide where to push: more qualified traffic, better engagement on key pages, smoother funnels, or stronger search visibility, but you will be choosing based on data that actually matters, not just whatever chart looks most impressive this week.

Need a quick summary of this article? Choose your favorite AI tool below: