60-Second Summary
This article catalogs every website metric worth tracking, explains what each one measures and conceals, and gives defensible ranges and baselines. It stresses that standard analytics count sessions not companies, so you need company-level identification to know whether visitors match your ICP and whether metrics should change decisions.
Key takeaways: Metrics fall into audience, acquisition, behavior and conversion families; use three-month baselines and trends rather than benchmarks, beware GA4 thresholds and measurement breaks, and remember none of these session-level numbers reveals which company was behind a visit.
Standout strategies & tactics: Report only the handful of KPIs that would change a decision, compare like periods and segments, pair complementary metrics (e.g., engagement rate with conversions or pages-per-session), and add reverse-IP company data to sort channels and pages by ICP fit.
Real-world lessons & frameworks: Use the KPI filter “would you change a decision?”, surface conversions per 1,000 sessions when traffic and conversion trends diverge, mark measurement-affected periods, and layer firmographics/company identification (e.g., Leadfeeder) to measure ICP ratio by channel, page and campaign.
*This summary was created with AI assistance, using our original content.
This article covers every metric worth tracking, how to judge each one, gives you a defensible range for each, and explains the one thing none of them can tell you: which companies were behind the sessions and whether those companies fit your ICP.
What counts as a website metric
A metric is any number your analytics tool reports: sessions, pageviews, conversion rate, engagement time. A KPI is the smaller set your team has decided to be judged on: the numbers tied to a goal someone will act on. Most teams track dozens of metrics but should only report the handful that would change a decision. If sessions drop 15% and nobody adjusts a campaign, sessions are a useful metric but not one of your KPIs.
This page covers what each metric counts, what it hides, and what a defensible range looks like. If you want the wider discipline and where metrics fit inside it, start with web analytics. If you came here looking for a one-paragraph definition rather than a full list, the web analytics glossary is the faster read.
Site performance and Core Web Vitals are a separate discipline with their own tooling and owner, so they are out of scope here.
The website metrics worth tracking
Each family below answers a different question. The tables share four columns so you can read them together.
Audience metrics
Audience metrics show how many people came and what they look like in aggregate.
Acquisition metrics
Acquisition metrics show where visitors came from and how they found you.
The three search rows above get their full treatment in seo website metrics.
Behavior metrics
Behavior metrics show what visitors did once they arrived.
Conversion metrics
Conversion metrics tell you whether visitors did what you wanted them to do. A conversion is the website action you define, like a form fill or a demo-page visit.
If you run an online store, orders, average order value and cart abandonment are a separate set, covered in ecommerce metrics.
The three metrics B2B teams over-report
Bounce rate, pageviews and time on page show up in almost every monthly deck, yet none of them tells you anything useful without the context around it.
In GA4, bounce rate is the share of sessions that weren't engaged: under 10 seconds, one page, no key event. It's simply 100% minus your engagement rate, so reporting both says the same thing twice. For a blog post that answers a question completely, a high bounce rate is the outcome you want. On a B2B site, a good google analytics bounce rate depends on page type, traffic source and what you defined as a conversion. A site-wide number means almost nothing. Pick one of the pair, usually engagement rate, and break it down by page type and channel.
Pageviews go up when a visitor clicks around lost, and they also go up when someone reads five articles in a row. The number alone can't tell you which scenario you're looking at, so reporting raw pageviews invites the wrong conclusion. Report pages per session segmented by channel and pair it with engagement time to separate genuine depth from confusion.
Time on page has a measurement problem. GA4 replaced it with average engagement time, which is more accurate, but neither tells you whether the visitor who spent four minutes on your pricing page is from an ICP-fit manufacturer or a student writing a paper. Report engagement time by page and pair it with conversion data to see whether attention is becoming action.
For the fuller version and the B2B lens on every metric, see b2b website metrics.
What every metric on this page hides
Every metric above counts sessions. None of them tells you which company the session belonged to.
In a Gartner survey of 646 B2B buyers, 67% said they prefer a buying experience with no sales rep involved (Gartner, 9 March 2026). Leadfeeder data shows roughly 2% of website visitors ever fill out a form.
GA4 also withholds rows with too few users when a report includes demographic or search query data, a mechanism Google calls "data thresholds." They protect privacy and can't be switched off, which means small B2B segments can vanish from reports entirely.
GA4 tells you a session came from a LinkedIn campaign, but it can't tell you that the session was a 200-person manufacturer in your ICP. That company-level view is the layer missing from every metric on this page.
Leadfeeder Web Insights adds that layer by showing which channels, campaigns and pages bring the companies you want, so you can spend where the leads actually come from. It sits alongside GA4 rather than replacing it. GA4 counts sessions and users, Leadfeeder counts companies, and the two sets of numbers won't match, because they count different things. Five people from one company are five users in GA4 and one company in Leadfeeder, and a visit only becomes a company once its IP address resolves to one. Install the tracking script and open the Overview after a week to see all visits charted against the share from ICP-fit companies.
Here's what each standard metric looks like when you add the company view:
For agency analysts building this report for a client, the ICP ratio answers the question that matters: is the traffic you're sending the right traffic? Sort the Traffic Sources report by ICP ratio to see which channels bring fit and which bring noise.
The mechanism is reverse IP lookup, which matches a visit against an IP-to-company database. It resolves a visit to an organization, not a person. Company-level data is where web analytics meets the broader practice of website intelligence, which looks at which organizations were behind the visits. To see the companies by name, use website visitor identification and filter by your target market.
One tracking script, ready in minutes. Install it, wait a week, then open the Pages report to see which pages are attracting ICP-fit companies and which are not.
How to pick the handful you actually report
Your website success metrics should be the handful that pass a simple filter: would you change a decision if this metric moved? If it goes up or down and nobody in the room would do anything differently, it doesn't belong in the deck.
Compare like periods on like segments. A month-over-month comparison that mixes organic growth with a one-off paid campaign hides the real trend. Break it out by channel and compare January to January rather than January to December, so seasonality doesn't masquerade as growth. When traffic and conversions move differently, report them as one rate, such as conversions per 1,000 sessions. Two lines on one chart invite a story the numbers don't support.
Every metric you put in front of leadership should come with a reason it moved. Before you look for a business reason, rule out a measurement one: a tag change, a consent banner update or a burst of bot traffic can move a number as much as real demand. Check the daily view, because a weekly or monthly chart spreads a two-day problem across the whole period. If a period is affected, mark it in the report rather than dropping it.
If sessions dropped 12% and you don't know why, say so. A number without context is worse than no number, because someone in the room will fill in their own story.
For the framework on choosing KPIs, see website kpis. For how to lay them out, see web analytics dashboard.
Frequently asked questions
What are the most important website metrics?
The most important website metrics are the ones tied to a decision you face. For most B2B marketing teams, that means sessions and users by channel, engagement rate by page type, conversion rate by source, and the share of identified companies that fit your ICP.
What is the difference between pageviews and sessions?
A pageview is a single page load, and a session is the entire visit that contains it. If a visitor lands on your homepage, clicks to your pricing page, then reads a case study, that's one session and three pageviews.
What is a good engagement rate for a website?
A good engagement rate depends on your traffic mix and content type. In a Siege Media study of 546 GA4 properties (August 2026), the median was 51% overall, with SaaS and B2B software sites averaging 46.6%. Your own three-month average is the most useful benchmark, because what counts as engaged varies by GA4 configuration.
How many metrics should a monthly website report include?
A monthly website report should include fewer metrics than you'd expect. Five to eight, each broken down by the segment that matters (channel, page type, or ICP fit), give you enough to explain what happened and decide what to do next. Every number should say which site, which dates and which metric it is.
Which website metrics actually predict revenue?
No single website metric predicts revenue on its own. The closest signal is conversion rate from ICP-fit companies, because it confirms the right accounts are visiting and acting. Pair it with your close rate and average deal size for a pipeline forecast.
Start the 14-day free trial, no credit card, and check the Pages report after a week. Terms and conditions apply.
