Top 20 Landing Page Metrics to Track in 2026

The landing page metrics to track are the ones that tell you what to change next, and most analytics dashboards show far more numbers than that. A typical report has sessions, page views, bounce rate, time on page, scroll depth and a dozen other figures, and none of them says what to do on Monday morning. Adding more metrics rarely fixes that. What helps is knowing which numbers measure the outcome you want, which ones explain why the outcome moved, and which ones are only context.

This guide picks 20 landing page metrics to track, groups them by where they sit in the visitor's path, and gives each one the same practical treatment: what it measures, how to read a high or low result, what can mislead you, and what to investigate next. It also covers how to set up measurement so the numbers can be trusted, how the metrics change by goal and traffic source, and how to put them into a dashboard that supports decisions.

What landing page metrics are, and which ones deserve to be KPIs

A landing page metric is any measurement of how people reach, use, and act on a page. Some of them are worth managing against a target. Many are only useful when something goes wrong. This section separates the terms so the rest of the guide is easier to use.

a. Metrics and KPIs: A KPI is a metric you manage against a goal.

Every KPI is a metric, but most metrics are not KPIs. A KPI is the small set of numbers that tell you whether the page is doing its job, such as the share of visitors who book a demo or the cost of each qualified lead. Scroll depth, for example, can be a useful metric without ever being a KPI. It helps you understand a page, and nobody should be rewarded for raising it.

b. Outcome and diagnostic metrics: One tells you what happened, the other suggests why.

Outcome metrics measure the result you care about: conversions, qualified leads, revenue. Diagnostic metrics sit upstream and help explain a change in the outcome: click-through rate on a button, form completion, load speed. When an outcome moves, you work backwards through the diagnostics to find where visitors stopped behaving the way you expected. That order matters. A diagnostic metric looked at on its own tends to send you off to fix things that were not broken.

c. Vanity metrics: Numbers that look good in a report and change no decision.

A vanity metric is one that rises or falls without altering anything you would do. Raw page views are the usual example. If traffic doubles because a broad ad campaign went live and conversions stay flat, the larger number tells you about spend and nothing about the page. Any metric can become a vanity metric if nobody acts on it, which is why this guide attaches a next step to each one.

d. Why there is no single dashboard: The right metrics depend on context.

The metrics worth your attention shift with the conversion goal, the traffic source, the audience, the funnel stage, the business model, the page type, the offer, and the cost of acquisition. A page selling a $20 product to cold social traffic and a page collecting demo requests from branded search should not share a dashboard. Treat the list below as a menu of landing page metrics to track, and treat the section on metrics by landing page goal as the guide to choosing.

A funnel framework for the landing page metrics to track

Sorting the landing page metrics to track by funnel stage keeps the list from becoming arbitrary. Each stage asks a different question, and each question points to a different set of numbers. This section gives the framework and the hierarchy that decides which metrics get your attention first.

a. Seven stages: Traffic, acquisition, engagement, page behaviour, conversion, quality and business outcome.

Each stage answers one question, and the 20 metrics below map onto them as shown.

Stage	The question it answers	Metrics in this guide
Traffic	Are people reaching the page?	1. Visitors and sessions
Acquisition	Where do they come from, and what did that traffic cost?	2. Traffic source and medium, 3. Ad click-through rate, 4. Cost per visitor, 5. New versus returning visitors
Engagement	Are visitors interacting with the page?	6. Engagement rate, 7. Average engagement time, 8. Scroll depth, 9. CTA click-through rate
Page behaviour	Where do they drop off or run into friction?	10. Form start rate, 11. Form completion rate, 12. Exit rate, 13. Page speed and Core Web Vitals, 14. Technical error rate
Conversion	Do they complete the intended action?	15. Landing page conversion rate, 16. Cost per conversion
Conversion quality	Are those conversions the right people?	17. Qualified lead rate, 18. Cost per qualified lead
Business outcome	Does the page contribute to revenue?	19. Revenue per visitor, 20. Return on ad spend

b. A three-tier hierarchy: Outcome, diagnostic, and supporting metrics.

Not all 20 deserve equal attention. Core outcome metrics measure whether the page achieves its business objective: conversion rate, qualified lead rate, cost per qualified lead, revenue per visitor, and return on ad spend. Diagnostic metrics explain movement in those outcomes: ad click-through rate, CTA click-through rate, form start and completion rates, page speed, technical errors, engagement rate and exit rate. Supporting metrics add context and should rarely drive a decision alone: visitor and session counts, new versus returning mix, engagement time and scroll depth. Which tier a metric falls into can change with the page. On a long-form page built to educate, engagement time may sit close to the core.

c. Start with the outcome and work backwards: This is the order that saves time.

When something looks wrong, begin at the outcome and ask which diagnostic explains it. If qualified leads fell, check whether conversions fell or whether quality fell. If conversions fell, check whether traffic changed, then whether the click-to-form path changed, then whether a technical problem appeared. Starting at scroll depth and working forward is slower, because nearly every diagnostic metric moves for several reasons.

Top 20 Landing Page Metrics to Track in 2026

Traffic metrics describe who arrives and what you paid for them. They matter mainly because they shape every number further down the page. A page that receives the wrong audience will look weak on engagement and conversion even if the page itself is sound. These five landing page metrics to track are useful mostly as context for the rest.

1. Visitors and sessions: Volume gives context and says little about performance.

• What it measures: How many people reached the page (visitors, which Google Analytics calls users) and how many visits they made (sessions). Page views count every load, including reloads, and on a single-page landing page they usually track sessions closely, so they add little on their own. This guide folds page views into this metric for that reason.
• How to read it: A rise can mean a campaign launched, a link spread, or a bot started crawling. A fall can mean budget ran out, a campaign paused, or tracking broke. Neither says anything about whether the page works. The useful move is to compare volume against the conversions it produced.
• What can mislead you: Sessions and visitors are not the same thing, so pick one and keep it consistent when you calculate rates. Bot and internal traffic inflate both. Consent choices can also reduce what your analytics sees, which a later section covers.
• What to do next: When volume changes sharply, look at source and medium before touching the page. If volume is steady and conversions move, the cause is more likely on the page or in the audience mix.

2. Traffic source and medium: The metric that changes how every other metric should be read.

Traffic source works as a way of splitting other numbers. Conversion rate, engagement, and cost all mean different things for a branded search visitor, a cold social visitor, and an email subscriber. Google's help page on campaign URLs lists five UTM parameters, of which source, medium and campaign are the ones to always include. It also notes that Analytics treats values as case-sensitive, so utm_source=google and utm_source=Google appear as separate rows.

• How to read it: Compare conversion rate, cost, and quality by source, then by campaign within a source. A source with a modest conversion rate can still be your best, if the people it sends become customers.
• What can mislead you: Inconsistent UTM naming splits one source into several. Untagged links fall into direct or other buckets. Attribution settings decide which touchpoint gets credit, so the same conversion can appear under different sources in different tools.
• What to do next: Agree on one naming convention before the next campaign launches, tag every paid and email link, and review source-level results before judging the page as a whole. Our guide on what organic traffic is and how to measure it covers how that source is usually reported.

3. Ad click-through rate: Whether the promise earns the click.

• What it measures: The share of ad impressions that produce a click. It is calculated as clicks divided by impressions. This is an ad metric that belongs here because it determines who arrives.
• How to read it: A low click-through rate may point to weak targeting, a dull ad, or an offer that doesn't interest the audience. A high one can mean a strong match, or an ad that promises more than the page delivers and attracts curious clicks. Google's own Quality Score documentation treats expected click-through rate, ad relevance, and landing page experience as three separate diagnostic inputs.
• What can mislead you: A rising click-through rate with a falling conversion rate often signals a message gap between ad and page. Click-through rate also varies by position, device, and match type, so averages across a whole account hide a lot.
• What to do next: When click-through is high, and conversion is low, put the ad and the first screen of the page side by side and check whether they make the same promise. Our article on what to do when ads get clicks but few conversions goes through that check.

4. Cost per visitor: What each arrival costs.

• What it measures: Ad spend divided by visitors, which for most paid campaigns is close to cost per click. It is the starting point of your campaign economics.
• How to read it: A high cost per visitor is not automatically bad. If those visitors convert at a high rate and the sale is worth a lot, the traffic can still pay for itself. A low cost per visitor is not automatically good either, because cheap traffic often carries lower intent.
• What can mislead you: Cost per visitor says nothing about what the visitor does next. Two campaigns with the same figure can differ completely in return. Also check that the visitors your ad platform reports and the visitors your analytics counts roughly agree, because they use different counting rules.
• What to do next: Read it alongside conversion rate. Cost per visitor divided by conversion rate gives you cost per conversion, which is where the decision actually sits.

5. New versus returning visitors: Familiarity changes behaviour.

• What it measures: The split between people seeing the page for the first time and people who have been before. It also appears as return visitor rate.
• How to read it: Returning visitors often convert differently from first-timers, since they have already seen the offer. A high return share on a short-cycle offer can suggest people need more than one visit to decide, or that retargeting is doing a lot of the work. Seline's guide frames a return visit as a sign of interest and a leading indicator, which is a reasonable hypothesis, though it is not a measured result.
• What can mislead you: Cookie deletion, device switching, and consent choices all break the link between visits, so returning visitors are usually undercounted. A returning visitor who converts may also be credited to a different source than the one that first persuaded them.
• What to do next: Compare conversion rate for the two groups before deciding that either needs a different page. If returning visitors convert far better, ask whether a first-visit page could answer the questions they needed a second visit to settle.

Engagement metrics: Are visitors interacting with the page?

Engagement metrics describe what visitors do on the page without yet asking whether they converted. They are diagnostic. None of them is valuable for its own sake, and each can mean opposite things depending on the page. This section shows how to read four of them without telling yourself a story the data does not support.

6. Engagement rate (and bounce rate): Read the definition your analytics tool actually uses.

Google Analytics 4 defines an engaged session as one that lasts longer than 10 seconds, has a key event, or has two or more page or screen views. The engagement rate is the percentage of engaged sessions, and bounce rate is the percentage of sessions that were not engaged, so the two always add to 100%.

• How to read it: A low engagement rate suggests that many visitors left quickly without doing anything the tool counts. That could mean poor relevance, a slow page, or a mismatched audience. It could also mean a page so clear that visitors converted somewhere else, or left to call you.
• What can mislead you: This is the metric most affected by outdated definitions. Several of the reference articles describe bounce rate as a visit with a single page view or a single request, which is how the older Universal Analytics worked. Under the GA4 definition, a visitor who stays 11 seconds counts as engaged even if they read nothing. Two tools can report the same page with very different bounce rates and both be correct. A very low bounce rate can also point to a tracking fault, such as an event firing on page load.
• What to do next: Segment by source and device before acting. A high bounce rate from one campaign points to targeting or message match. A high bounce rate only on mobile points to experience or speed. Do not set a company-wide target for it.

7. Average engagement time: Time on page tells several stories.

• What it measures: How long a page was in focus while the visitor was using it. Google describes user engagement as the amount of time someone spends with the web page in focus, and Google Analytics uses it to populate engagement time metrics.
• How to read it: A long time can mean strong interest, complex content, confusion, or difficulty finding the next step. A short time can mean a fast conversion, immediate rejection, poor relevance, or an obvious page that needed no study. The number alone cannot tell you which. Measureschool makes the same point about long times on pages with broken forms or buttons.
• What can mislead you: Averages are pulled around by a few very long sessions, and a background tab left open can distort them. Time is also measured differently in different tools, so a number from one platform should not be compared with another.
• What to do next: Pair it with scroll depth and conversion rate. Long time plus low conversion plus low scroll suggests people are stuck near the top. Short time plus high conversion suggests an efficient page. Session recordings can confirm which story applies. Microsoft Clarity, for example, offers click maps, scroll maps, and recordings.

8. Scroll depth: How far down the page people read.

• What it measures: The furthest point on the page a visitor reached, usually reported in milestones such as 25%, 50%, 75% and 100%. In Google Analytics 4, the default scroll event fires the first time a user reaches 90% vertical depth, so it records one threshold and not a distribution. Finer milestones need custom setup.
• How to read it: If most visitors never reach the section with the offer details or the form, that section may be placed too low, or the top of the page may fail to hold attention. If nearly everyone reaches the bottom but conversion is low, the page is being read, and the problem is likely the offer, the trust signals, or the form.
• What can mislead you: Scroll depth measures how far someone got and says nothing certain about attention. Someone can scroll to 90% in a second. On a short page, nearly everyone hits the milestone, which makes the metric useless there. Pages with accordions or tabs hide content that scroll depth cannot see.
• What to do next: Check where your key proof, pricing, and call to action sit relative to where people stop. Move the element that matters, then see whether click-through or conversion shifts. Use it as a placement guide.

9. CTA click-through rate: Whether the main action gets noticed and chosen.

• What it measures: The share of visitors (or sessions) who click the primary call to action. The formula is CTA clicks divided by visitors who could see the button, multiplied by 100.
• How to read it: A low rate can point to a weak offer, an unclear button label, poor placement, a mismatch with intent, or a button that looks inactive. A high rate with low form completion often means the click promised something the next step did not deliver, which belongs to the next section.
• What can mislead you: If the page has several calls to action, a single rate hides which one works. A click on a button that opens a form and a click on a button that leaves the page are different things. A sudden drop to near zero is usually a broken button and rarely a change in behaviour, so check the page before drawing any conclusions about the audience.
• What to do next: Track each CTA as its own event with a clear name, and compare clicks by placement and by source. If the primary CTA is clicked often but conversion stays low, move your attention to the form and the step after the click.

Page behaviour metrics: Where do visitors drop off or hit friction?

Behaviour metrics follow visitors through the steps between arriving and converting. They are the most useful diagnostics on a page, because a drop between two steps points to the part of the page worth investigating. This section starts with a simple funnel and then covers the five landing page metrics to track that locate friction.

a. A hypothetical funnel: Visit, engage, CTA, form start, form complete, qualified lead, customer.

The numbers below are invented to show how to read drops between steps. They are not benchmarks.

Step	Count	Share of the previous step
Visits	2,000	n/a
Engaged sessions	1,000	50%
CTA clicks	300	30%
Form starts	150	50%
Form completions	90	60%
Qualified leads	36	40%
Customers	9	25%

The biggest percentage drop is between engagement and the CTA click, but that does not make it the first thing to fix. Only half of CTA clickers start the form, and only 60% of starters finish, so the form path loses a lot of people who had already shown intent. Visitors who have clicked are warmer than visitors who merely scrolled, so a point recovered further down the funnel may be cheaper to win than a point recovered near the top. Treat that as a hypothesis to test. The drop between completion and qualified lead belongs to a different team, and it may reflect the offer or the audience more than the page.

10. Form start rate: Whether the form looks worth beginning.

• What it measures: The share of visitors who interact with the form at all. Google Analytics 4's enhanced measurement can record a form_start event the first time a user interacts with a form in a session, and a form_submit event when the form is submitted. Divide form starts by visitors, or by the visitors who reached the form, depending on the question you are asking.
• How to read it: A low rate suggests visitors are not convinced enough to begin. Candidates include the offer, the headline, how much the form seems to ask for, and trust. A healthy start rate with a poor completion rate moves the problem inside the form.
• What can mislead you: Automatic form events depend on how your form is built. Forms that load in a frame, load late, or use custom scripts may not be detected, so the event can be missing or double-counted. Check it against a real test submission.
• What to do next: Look at what visitors see just before the form. If the form appears only after a long scroll, check where scroll depth drops. If it is visible at once and still ignored, test the framing of the offer before the field count.

11. Form completion rate (and abandonment): Where intent turns into friction.

• What it measures: The share of people who start the form and submit it. Form abandonment rate is the other side of the same number, so there is no need to report both. If 150 people start and 90 finish, completion is 60% and abandonment is 40%.
• How to read it: A low completion rate is the clearest sign of form friction: too many fields, confusing questions, an error message that doesn't explain itself, a request for information people don't want to give, or a lack of trust at the moment of sharing details. It can also reflect a mismatch in expectations if the click promised a quick download and the form asks for a phone number.
• What can mislead you: A longer form may deliver fewer submissions but better leads, so a lower completion rate is not automatically a loss. Shortening the form can raise completions and lower quality. Tools that track field-level drop-off can show which question people abandon at, which is more useful than the overall rate.
• What to do next: Find the field where people stop, test removing or reframing it, and watch qualified lead rate afterwards. Never judge a form change on completion rate alone.

12. Exit rate: Where people leave, which is not always a problem.

• What it measures: The share of views of a page that were the last in a session. Measureschool notes that GA4 reports exits as a metric but not a ready-made exit rate, so you may have to calculate it in a spreadsheet or reporting tool.
• How to read it: On a single-page campaign page with no further steps, nearly every session ends there, so exit rate carries almost no information. It becomes useful on multi-step flows, where a high exit rate on step two says that step is losing people. Seline points out that exits from a thank-you page after a form submission are normal.
• What can mislead you: A high exit rate from a page that is meant to be a destination is expected. It also differs from bounce rate, which counts only sessions that were not engaged.
• What to do next: Use it to locate the leaving point in a multi-step path, then use recordings or form analytics to find out why. Skip it for one-page campaign pages.

13. Page speed and Core Web Vitals: Technical friction that copy testing cannot fix.

• What it measures: How quickly the page loads and responds. Google's Core Web Vitals are Largest Contentful Paint for loading, Interaction to Next Paint for responsiveness, and Cumulative Layout Shift for visual stability. Google's web.dev documentation gives the good thresholds as 2.5 seconds or less for LCP, 200 milliseconds or less for INP, and 0.1 or less for CLS, assessed at the 75th percentile of real page loads and separately for mobile and desktop.
• How to read it: A page that fails on mobile may be losing a large share of its paid traffic before the headline loads. If speed is poor, there is little point running another copy test, because visitors who never see the page cannot respond to it. Layout shift matters too, since a button that jumps as the page loads can cause a misclick.
• What can mislead you: Lab tools and field data differ. Field data comes from real users and captures the variety of devices and networks, while lab tools run controlled tests and cannot measure INP directly. A fast result on your office connection says little about visitors on a mid-range phone.
• What to do next: Check field data for the landing page itself, split by device. If a threshold fails, fix that before spending time on layout or copy tests. Heavy images, third-party scripts, and late-loading embeds are common places to look.

14. Technical error rate: The metric that explains conversion drops that make no sense.

• What it measures: How often something breaks. No platform reports one standard number under this name, so you define it. A useful version is failed form submissions divided by submission attempts, plus the count of tracking events that stopped firing.
• How to read it: Errors rarely announce themselves. A form that returns an error on certain phone number formats, a button that fails in one browser, or a conversion tag that stops firing after a site update can all look like a sudden change in audience behaviour. If conversion rate falls sharply overnight with no change in traffic, suspect a break before suspecting people.
• What can mislead you: Silent failures are the danger. A broken event can read as zero conversions, and a duplicated event can read as a surge. Neither says anything about how visitors feel about your offer.
• What to do next: Test the full path on a phone and a computer after every release, confirm the conversion shows up in analytics, the ad platform, and your CRM, and set an alert for conversions dropping to zero. Our guide to diagnosing a landing page that gets traffic but no conversions lists the technical checks to run first.

Conversion metrics: Do visitors complete the intended action?

Conversion is where the page either does its job or does not. Of all the landing page metrics to track, it needs the most careful definition, because a vague conversion event makes every rate downstream unreliable. This section covers conversion rate in depth and the cost per conversion that goes with it.

15. Landing page conversion rate: The central outcome metric, and the easiest to define badly.

• What it measures: The share of visitors who complete the action the page exists for. The basic formula is conversions divided by visitors, multiplied by 100. Google Ads defines conversion rate as conversions per ad interaction, so 50 conversions from 1,000 interactions is 5%. The same page also says that when you count every conversion rather than one per click, a rate can exceed 100%.
• What counts as a conversion: Decide this before you read any rate. The event should be the outcome you want, such as a submitted form, a completed purchase or a booked slot, and not an earlier step such as a button click. Google Analytics describes a key event as an event that measures an action that is particularly important to the success of your business. Treat clicks and scrolls as micro-conversions and keep them separate from the primary one.
• The denominator problem: Ad platforms usually divide by clicks or interactions. Analytics tools divide by sessions or users. The same page can show three different conversion rates in three tools, and all three can be correct. Pick one denominator for reporting and label it. Measureschool notes that GA4 offers both session-based and user-based rates in its Explorations reports.
• Primary and secondary conversions: In Google Ads, primary conversion actions appear in the Conversions column and can be used for bidding, while secondary actions appear in All conversions and are for observation. Keep one primary action per page. If every micro-action counts as a conversion, the rate will look healthy while the business outcome stays flat.

Conversion type	A sensible primary event	The common trap
Purchase	Completed order	Counting add-to-cart as the conversion
Lead	Submitted form that passes validation	Counting spam and test submissions
Demo request	Confirmed request	Ignoring whether the lead is qualified
Free trial or signup	Account created	Stopping at signup when activation is the goal
Consultation or booking	Confirmed time slot	Counting clicks on the calendar link
Download	Completed form and delivery	Treating low-intent downloaders as leads
Quote or contact request	Submitted request	Counting phone-number clicks and form fills as the same thing

• How to read it: A low rate can point to poor traffic quality, a weak offer, message mismatch, friction, or a technical break. A high rate can mean a strong page, but also a very easy conversion that attracts people who won't buy. Conversion rate gets harder to interpret as soon as the conversion is cheap to give.
• What can mislead you: There is no universal good rate. The Unbounce Conversion Benchmark Report analysed more than 41,000 pages and 464 million visitors from July 2023 to July 2024, and reported a median of 6.6% overall, with industry medians that ranged from 3.8% for SaaS to 12.3% for entertainment. Its methodology does not define how conversion was calculated, and it excludes pages with fewer than 50 visitors. Use figures like these as a rough reference for your own history, not as a target. Our article on what a good landing page conversion rate looks like goes through why.
• What to do next: Read the rate by source, device, and campaign before deciding the page is the problem. Then look at the diagnostics above it in the funnel: CTA click-through, form completion, and speed. Our guide on what conversion rate is and how to measure and improve it covers the basics in more depth. When you test a change, fix the sample size in advance and avoid checking results early, which raises the chance of a false winner.

16. Cost per conversion: What each outcome costs, which is not the same as what it is worth.

• What it measures: Ad spend divided by conversions. Google Ads shows it as Cost / conv., calculated as total cost divided by the conversions column. If $2,000 produces 40 conversions, cost per conversion is $50. Marketers also call it cost per acquisition, or CPA, and cost per lead when the conversion is a lead.
• How to read it: The figure only means something against the value of a conversion. Seline makes the useful point that acceptable cost depends on profit per sale, and no universal target exists. A $50 conversion is cheap for a $2,000 contract and expensive for a $15 purchase. A falling cost per conversion can result from better targeting, a better page, or a cheaper but lower-quality audience.
• What can mislead you: It treats every conversion as equal. A campaign can lower its cost per conversion while the leads get worse. Platform attribution and conversion modelling also change the denominator, and one reference article writes the formula as CPC, which normally means cost per click.
• What to do next: Compare it with the value of a conversion and with cost per qualified lead. If it rises, check whether cost per visitor rose or conversion rate fell, because the fixes are different.

Conversion quality and business outcome: Did the right people convert, and did it pay?

A landing page is not successful because it generates more form submissions. It succeeds when the conversions turn into something the business values. These four landing page metrics to track sit mostly outside the page, and they are the ones most dashboards leave out. This section also separates the page's performance from the economics of the campaign that feeds it.

17. Qualified lead rate: The share of conversions your sales team would accept.

• What it measures: Qualified leads divided by total leads from the page. How you define "qualified" is the whole question. It might be a marketing-qualified lead, a sales-accepted lead, or a lead that became an opportunity. Pick a stage, write the criteria down, and stick to them. The metric can be extended downstream to opportunity rate and lead-to-customer rate.
• How to read it: A low rate with a high conversion rate points to a conversion that is too easy, an audience that is too broad, or an offer that attracts the wrong people. A high rate with low volume may mean the page is filtering well and could afford to loosen up.
• What can mislead you: Qualification depends on sales follow-up. If leads are contacted slowly or inconsistently, a good page can look weak. Qualification also lags, sometimes by weeks, so recent campaigns will always look worse than older ones.
• What to do next: Break the rate down by source, campaign, and form. Share the breakdown with sales and ask why leads from the worst segments were rejected. Google Ads lets you import offline conversions, including enhanced conversions for leads and imports matched by click ID, so the platform can learn from qualified leads as well as form fills. Google reports a median 10% increase in conversions for advertisers that combined first-party data with click IDs, which is Google's own figure.

18. Cost per qualified lead: The price of the leads you actually wanted.

• What it measures: Ad spend divided by qualified leads. If 40 conversions cost $2,000 and 12 pass qualification, cost per conversion is $50 and cost per qualified lead is about $167.
• How to read it: It is usually a better guide than cost per conversion for lead generation, because it penalises campaigns that buy cheap leads nobody wants. Customer acquisition cost is its more demanding sibling, which divides the total sales and marketing cost by new customers. Decide which costs you include, because teams define it differently.
• What can mislead you: It needs CRM data, and small numbers make it volatile. A single week with three qualified leads can swing the figure wildly.
• What to do next: Compare it by campaign and source over a long enough window to have a stable count. Use it to decide where to shift budget, and compare it with what a qualified lead is worth to the business.

19. Revenue per visitor: What each arrival is worth.

• What it measures: Revenue divided by visitors. It combines conversion rate and order value into one number, which is why ecommerce teams like it. For lead generation, use pipeline value or expected value per visitor, if your data supports it.
• How to read it: A page can raise conversion rate by pushing discounts while lowering revenue per visitor. Revenue per visitor catches that. Comparing it with cost per visitor shows whether each arrival returns more than it costs.
• What can mislead you: Revenue arrives at different speeds. For long sales cycles, current revenue understates a page's effect. A few large orders can also dominate a small sample.
• What to do next: When testing page variants for a sales page, treat revenue per visitor as the deciding metric alongside conversion rate, and give the test enough time and traffic to show a stable difference.

20. Return on ad spend: Revenue earned per dollar of ad spend, and a campaign metric before it is a page metric.

• What it measures: Conversion value divided by ad cost. Google Ads reports this as Conv. value/cost, calculated as total conversion value divided by total cost of all ad interactions. A ROAS of 4 means $4 of conversion value for each $1 spent.
• How to read it: It tells you whether a campaign returned more than it cost in revenue terms, but revenue is not profit. A ROAS of 3 is excellent on a high-margin product and a loss on a thin-margin one. A low ROAS may reflect the page, but also the audience, the bid strategy, the product price, or the sales cycle.
• What can mislead you: Platform ROAS depends on the conversion values you send and on the attribution model. Different platforms credit the same sale differently, so summing their ROAS figures overstates results. Conversion modelling and consent choices also change what is counted.
• What to do next: Compare ROAS across campaigns that share a goal and margin, and use cost per qualified lead where revenue is delayed. Treat a page change as the likely cause only after you've ruled out budget, bidding, and audience changes.

a. Page performance versus campaign economics: A strong page can sit inside a weak campaign.

Conversion rate describes the page. Cost per conversion, cost per qualified lead, and ROAS describe the page plus the cost of the traffic plus the value of the conversion. The two can disagree. Here is a hypothetical comparison.

	Campaign X	Campaign Y
Spend	$5,000	$2,000
Visitors	1,000	500
Conversions	100	15
Conversion rate	10%	3%
Value per conversion	$30	$400
Cost per conversion	$50	about $133
Revenue	$3,000	$6,000
ROAS	0.6	3

Campaign X has the stronger page by conversion rate and loses money. Campaign Y converts fewer visitors and returns three times its spend. If you judged only by conversion rate, you would back the wrong campaign. If you judged only by ROAS, you might miss that Campaign Y's page has room to improve. Both views are needed, and they answer different questions.

Which landing page metrics to track for each goal

The landing page metrics to track change in priority with the page's purpose, so the same 20 rank differently depending on what the page is for. This section shows how the hierarchy shifts across five common goals, so you can choose a short list and leave the rest as context. The aim is to give each goal two or three core metrics, a few diagnostics, and a clear idea of what to ignore.

Page goal	Core outcome metrics	Main diagnostics	Often safe to treat as context
Lead generation	Conversion rate, qualified lead rate, cost per qualified lead	Form start and completion, CTA click-through, speed	Scroll depth, new versus returning
Ecommerce	Conversion rate, revenue per visitor, ROAS	Product engagement, add-to-cart behaviour, speed on mobile	Visitor counts, time on page
SaaS demo	Demo conversion rate, qualified opportunity rate, pipeline per visitor	Form completion, CTA click-through, source quality	Exit rate, scroll depth
Free trial	Signup rate, activation rate, paid conversion	Form completion, error rate, source quality	Engagement time
Content download	Download conversion, lead quality, later engagement	Form completion, scroll depth, CTA click-through	Revenue per visitor in the short term

a. Lead generation: Judge the page by the leads that survive qualification.

For a service business or B2B offer, conversion rate and cost per lead are only the first read. Qualified lead rate and cost per qualified lead decide whether the page is working. Form start and completion rates tell you where to look when conversion falls.

b. Ecommerce: Revenue per visitor catches what conversion rate misses.

Because order value varies, conversion rate alone can mislead. Revenue per visitor and ROAS bring in value. The diagnostics are product engagement and add-to-cart behaviour, which are the micro-conversions that precede a purchase, along with mobile speed.

c. SaaS demo pages: Pipeline matters more than requests.

Demo requests are cheap to generate and expensive to handle. The metrics that matter move downstream: how many requests become qualified opportunities and how much pipeline each visitor creates. Form completion and source quality explain much of the variation.

d. Free-trial pages: A signup is where the outcome begins.

A signup that never activates costs money to acquire and returns nothing. Track activation, meaning whether the new user reached the first meaningful action in the product, and paid conversion, in addition to signup rate. Which action counts as activation is a product decision, so define it with the product team.

e. Content download pages: Check what happens after the download.

A download is a low-commitment conversion. It is only worth counting if downloaders go on to engage, reply, or move down the funnel. Lead quality, and later engagement matter more here than the download rate itself.

How traffic source changes what the metrics mean

A conversion rate of the same size can be a success from one channel and a failure from another. Intent, familiarity with your brand, cost, and funnel stage all differ by source, so benchmark each source against itself and against your own history. The landing page metrics to track for paid search are not automatically the ones to watch for paid social. This section runs through nine sources with the points that change interpretation. Official platform documentation for Microsoft Advertising, Meta, and LinkedIn could not be opened during research, so the notes on those platforms are practical reasoning with no documentation behind them.

a. Google search ads: People are asking for something, so message match is the test.

Searchers state intent in the query, which usually makes search traffic the highest-intent paid source. Expect higher click-through and conversion rates than from interruption channels, and expect poor message match to show up quickly. Google's help page on landing page experience recommends matching the page to the ad and keywords. Watch search terms as well as keywords, since irrelevant queries lower conversion rate without any fault in the page.

b. Microsoft ads: Similar intent, a different audience.

Search intent works the same way, but the audience mix, device split, and volume differ from Google's, so do not assume results will transfer. Track it as its own source, and tag it separately. Low volume makes rates noisy, so wait for a decent sample before reacting.

c. Meta ads: Interruption traffic, so expect lower intent and more scrolling.

People on social platforms did not come looking for you. That tends to mean lower conversion rates, shorter visits, and more mobile traffic. Judge these campaigns on cost per qualified lead or ROAS, not on bounce rate, and test whether a softer first offer works better than a direct request.

d. LinkedIn ads: Narrow, expensive audiences where lead quality is the point.

LinkedIn targets by job attributes, so cost per click tends to be high and the audience small. Quality metrics matter more than volume. LinkedIn Lead Gen Forms prefill profile data, which can reduce friction but may also change who completes the form, so compare lead quality against your landing page form.

e. Organic search: Intent varies by query, so segment by landing query.

Organic visitors range from people who have never heard of you to people searching your brand. Conversion rate for a page depends heavily on the query that brought them. Look at the pages and queries separately rather than reading one organic average.

f. Email: Warm audiences convert at different rates from cold ones.

Subscribers already know you. Expect higher engagement and conversion, and check whether the list is segmented. The same email sent to your entire list and to recent buyers will produce different pictures of the same page.

g. Referral: The quality depends on who sends it.

A mention in a trusted publication and a link in a directory behave differently. Look at referrals by referring site, and be careful about drawing conclusions from small counts.

h. Display: Low intent and high reach.

Display traffic reaches many people who were not looking, and clicks can include accidental ones. Engagement rate and conversion rate are typically lower, and view-through effects are hard to measure. Treat it as an awareness channel and judge it on assisted effects and downstream quality.

i. Retargeting: High conversion rates that mostly reflect who you chose.

People who have already visited are likely to convert at a higher rate than new visitors, but that is partly because you picked people who showed interest. Retargeting can make an account look better than the underlying acquisition is. Report it separately from prospecting, and compare it against the first-touch campaigns that fed the audience.

Segment before you conclude: Aggregate numbers hide the problem

An average across all visitors can hide a page that works well for one group and badly for another. A page can look healthy overall while mobile visitors have a very different experience. This section lists the cuts worth making on any of the landing page metrics to track, with the question each one answers.

a. Device: Mobile and desktop visitors use the page differently.

Compare conversion rate, form completion, and speed by device. A gap points to layout, form usability, or speed on one of them. Core Web Vitals are assessed separately by device for this reason. The Unbounce report's own data shows that most visits were on mobile and that desktop converted somewhat better overall, but that is an aggregate and should not stand in for your own split.

b. Campaign, ad group and keyword: Find which traffic carries the result.

A page serving ten campaigns has ten different audiences. Segment by campaign and, for search, by keyword and search term. Our guide on how many landing pages a campaign needs covers when the answer is to split the page as well as the report.

c. Audience, geography and new versus returning: Check whether one group is dragging the average.

Different audiences may need different offers. A region with a different language, price sensitivity, or purchase habit will convert differently. Compare new and returning visitors before deciding whether either group needs a new page.

d. Landing page variant: Compare variants on the same traffic.

When you test, split the same traffic between versions and compare on the same conversion definition. For Google Ads custom experiments, Google notes that experiments share traffic and budget. Avoid comparing a variant that ran last month with one that ran this month, since the audience and season will have changed.

e. Funnel stage: Watch each step along the way.

A flat overall conversion rate can hide one step improving and another getting worse. Report the stage-to-stage rates from the funnel earlier so changes show up where they happen.

Metrics you should not read alone

Some landing page metrics to track are misleading when viewed in isolation, and the danger is greatest when they're the only number in a report. Each one below has a pairing that makes it safer. This is a short table because the point is practical: before acting on a number, look at the number next to it.

Metric	The risk of reading it alone	Pair it with
Traffic or page views	Rewards volume over fit	Conversion rate and source
Time on page	Long can mean interest or confusion	Scroll depth, conversion rate, recordings
Bounce rate	Depends on the tool's definition	Source, device, conversion rate
Scroll depth	Shows reach with no sign of attention	CTA click-through, engagement time
Ad click-through rate	Can reward misleading promises	Landing page conversion rate
Conversion rate	Easy conversions inflate it	Qualified lead rate and cost per conversion
Cost per conversion	Cheap conversions can be worthless	Cost per qualified lead, revenue per visitor
ROAS	Revenue is not profit	Margin, cost per customer, time lag

Metric relationships: Treat each pairing as a hypothesis

Landing page metrics to track become more useful in pairs. Each combination below suggests where to look first. These are hypotheses to check, and none of them is a guaranteed explanation. The same pattern can come from more than one cause, and sometimes from several at once.

a. High traffic and low conversion rate: Look at who arrives before what they see.

Check traffic quality and intent first, then message match between the ad and the page, then the offer and the page experience. Broad targeting is a frequent cause, and a rewrite of the page will not fix it.

b. Low traffic and high conversion rate: The page may be fine and the reach too small.

Look at acquisition, budget, targeting, and reach. A small, well-matched audience converts well and then runs out. The question becomes whether you can reach similar people at a similar cost.

c. High conversion rate and low lead quality: The conversion may be too easy.

Check the offer, the qualification questions on the form, the audience, and what the messaging promises. The page may be attracting people who want the free thing and not the product.

d. High CTA clicks and low form completion: Look at what happens after the click.

Investigate form length, field types, trust at the point of asking, technical errors, and whether the form matches what the click implied.

e. High conversion rate and poor revenue: Follow the leads past the form.

Check customer quality, product economics, the offer, attribution, and downstream conversion. The page might be producing conversions the business can't monetise, or revenue might be arriving later than the report window.

Our article on what to do when ads get clicks but few conversions applies several of these patterns to a paid campaign.

Benchmarks: What to use and what to ignore

Benchmarks are tempting because they give a number to aim at. Most circulating figures don't say where they came from, which makes them close to useless for decisions.

a. What the reference articles cite: Mostly unsourced averages.

The six reference articles quote figures such as a 2% to 5% typical conversion rate, an average of 2.35%, and bounce rates of around 40% to 45%, usually without a date, industry, traffic source, or definition. One cites a 2014 URL for its conversion benchmark. Treat figures like these as folklore until the source is visible.

b. Benchmarks with disclosed methods: Useful as context.

Source	Scope	Figure	What limits it
Unbounce Conversion Benchmark Report	41,000+ pages, 464M visitors, July 2023 to July 2024, machine-learning topic classification	Median 6.6%; SaaS 3.8%, ecommerce 4.2%, legal 6.3%	Conversion calculation not defined; pages under 50 visitors excluded
WordStream Google Ads benchmarks	14,197 US accounts, August 2017 to January 2018	Search 3.75%, display 0.77%	Old; conversions divided by clicks; based on medians

c. How to use a benchmark: Compare against your own history first.

Your own trend, by source and by device, is more reliable than any external figure. If you use a benchmark, check that the page type, industry, traffic source, conversion definition, and geography match yours. If they don't, the comparison is not meaningful. Our article on what a good landing page conversion rate is explains this in more detail.

Set up measurement so the numbers can be trusted

Landing page metrics to track are only as good as the tracking under them. A misnamed event or a duplicated tag can send you in the wrong direction for months. This section covers the setup that matters most, based on Google's documentation where it applies.

a. Conversion events: Define the outcome before you build tracking.

Mark the one event that represents success as a key event in Google Analytics. Google notes that you can turn a key event into a Google Ads conversion. Keep micro-actions as separate events, and keep them secondary in the ad platform.

b. Event naming: Use a consistent, readable convention.

Name events by what happened, in the same style every time, and document them. For example, a form submit event and a CTA click event with a parameter naming which button. Consistency makes it possible to compare pages later and avoids duplicated or orphaned events.

c. UTM parameters: Tag every paid and email link the same way.

Use source, medium, and campaign on every link, and decide on lowercase values. Because values are case-sensitive, inconsistent capitalisation fragments your reports.

d. Form and CTA tracking: Check what automatic events miss.

Google Analytics 4's enhanced measurement can capture scroll, outbound clicks, and form interactions, but only under particular conditions, and the scroll event fires once at 90%. Verify each with a real test, and add custom events where the automatic ones don't fit your page. Recommended events should be sent with their prescribed parameters, and Google suggests DebugView for checking setup.

e. Duplicate prevention: Count each lead once.

Google Ads lets you count one conversion or every conversion per click, and suggests one for leads and every for sales. A thank-you page tag that fires twice, or a page that can be reloaded, will inflate leads. Test for it.

f. CRM and offline conversions: Connect the page to what happened next.

Pass the source and campaign into the CRM with each lead, so qualified lead rate and cost per qualified lead can be calculated by campaign. Import qualified or closed outcomes back into the ad platform where it supports this, using the offline conversion import options.

g. Consent and privacy: Know what you are not seeing.

Google's Consent Mode lets tags adjust behaviour according to visitor consent. When consent is denied, tags send cookieless pings, and Google fills gaps with behavioral and conversion modelling, but only under an advanced implementation. Under a basic one, you do not get modeled data in Google Analytics. That means your counts may be lower than reality in regions with strict consent rules, and different tools will disagree more. Take legal advice on consent in your regions.

h. Attribution: Platform numbers are estimates.

Google Ads uses data-driven attribution as the default for most conversion actions, and changing the model affects only future conversions. Your analytics, ad platform, and CRM will not match exactly. Use them to compare options and favour the source closest to revenue when they conflict. Assisted conversions, which credit earlier touchpoints, are worth a look when a channel looks weak on last-click results but feeds others.

A dashboard that supports decisions: Four views that each answer one question

A dashboard with 20 numbers and no decisions is a poster. The way out is to give each view one audience and one question, and to attach to each number a decision it can trigger. This section proposes four views that organise the landing page metrics to track, built around the outcome-first order from earlier.

a. Executive view: Is the page paying for itself?

Show conversion rate, cost per conversion, qualified lead rate, cost per qualified lead, revenue or pipeline per visitor, and ROAS. Keep it to one screen with trends over time. The decision it informs is where to put budget.

b. Acquisition view: Is the right traffic arriving?

Show visitors by source and medium, ad click-through rate, cost per visitor, and the new versus returning mix. The decision is which campaigns to scale, fix, or pause.

c. Behaviour view: Where do visitors lose momentum?

Show engagement rate, CTA click-through, form start rate, form completion rate, and for multi-step flows, exit rate. The decision is which part of the page to test next.

d. Diagnostic view: What explains a change?

Show conversion rate and qualified lead rate split by device, campaign, audience and page variant, plus page speed and technical errors. This view is for investigation and is too detailed for a weekly meeting. Our article on the build, measure, and improve loop for campaign pages describes how teams turn these reviews into a routine.

Turning metrics into actions: What a number may indicate and what to do about it

A metric earns its place among the landing page metrics to track by changing what you do. The table below connects common signals to possible meanings, what to investigate, and a reasonable first action. None of the rows proves a cause. Each is a starting point for investigation.

Signal	What it may indicate	Investigate	Possible action
Low conversion rate	Weak post-click performance, wrong traffic, or a tracking fault	Intent, offer, message match, page experience, tracking	Test the highest-impact hypothesis
Rising ad CTR, falling conversion rate	Ad and page promise different things	Ad copy next to the first screen	Align the headline and offer with the ad
Low CTA click-through	Unclear offer, weak placement, or low intent	Button label, position, relevance, source	Test CTA copy and placement
High CTA clicks, low form completion	Form friction or broken expectations	Fields, errors, trust, form length	Simplify or redesign the form
Low engagement rate on mobile only	Mobile experience or speed problem	Core Web Vitals, layout, form usability	Fix speed and layout for mobile
Conversion drops to near zero overnight	Broken tag or form	Test submission, tag firing, deploy history	Fix the break and restore tracking
High conversion, poor lead quality	Weak qualification or wrong audience	Audience, offer, form questions, source	Improve qualification or targeting
Low cost per conversion, low revenue	Cheap but low-value conversions	Cost per qualified lead, customer quality	Optimise toward qualified outcomes
Strong ROAS, thin volume	Narrow audience converting well	Budget, reach, targeting	Test expansion carefully

Common measurement mistakes

Most mistakes in choosing and reading landing page metrics to track come from a small set of habits. This section groups the common ones so you can check your own setup against them.

a. Tracking everything with no decision framework: More metrics, same confusion.

If a metric has no action attached to it, remove it from the regular report. Keep it available for investigation.

b. Treating every metric as a KPI: Targets on diagnostics distort behaviour.

Setting a target for scroll depth or time on page invites changes that raise the number without helping the business. Keep targets for outcome metrics.

c. Optimising for traffic volume or conversion rate alone: Both reward the wrong thing eventually.

Volume rewards broad targeting. Conversion rate alone rewards easy conversions. Pair each with a quality or economics metric.

d. Ignoring lead quality: The most expensive blind spot in lead generation.

Without feedback from sales, a campaign can improve its cost per lead while its results get worse.

e. Inconsistent conversion definitions and incomparable comparisons: Compare like with like.

Changing what counts as a conversion mid-quarter breaks every trend. Comparing a retargeting page with a cold-traffic page, or a demo page with an ebook page, produces differences that have nothing to do with page quality.

f. Ignoring device and campaign intent: Averages conceal them.

Always check the split before concluding that a page needs a rewrite.

g. Outdated analytics definitions: Check which bounce rate you are reading.

Several of the reference articles still describe bounce rate using older definitions and point to screenshots from earlier analytics versions. Check the current documentation for your tool.

h. Benchmarks without context and correlation as causation: Both produce confident, wrong decisions.

A benchmark with unknown scope is an anecdote. A metric that moved after a change was made may have moved for other reasons, such as a campaign, a season, or a tracking update.

i. Decisions from tiny datasets and ignored tracking errors: Wait for the evidence.

Rates from a few dozen visitors swing widely. Work out how much traffic a test needs before it starts, and do not stop it the moment it looks good. Check that the data is complete before analysing it.

j. Reporting without actions and dashboards nobody uses: The test is whether anyone changes something.

If a report has been sent for months and nothing has changed because of it, the report is the thing to fix.

AI-assisted landing page analytics: AI can find patterns, and you still have to establish causes

AI tools can speed up the work of reading landing page metrics to track, and they can mislead in a confident voice. The useful division is between finding patterns, where AI is fast, and establishing causes, where it needs your judgement and evidence. This section covers where it helps and where to hold back.

a. Where it helps: Summaries, segments and questions.

AI can summarise a large export, flag unusual spikes or drops, split performance by segment, compare campaigns, and turn a dashboard into a list of questions to investigate. It can group open-text survey or chat feedback into themes, suggest experiment ideas from a funnel drop, and draft recurring reports. Each of these saves time without making a decision for you.

b. Where to be careful: Causation, small samples, and bad data.

An AI tool that says conversion fell because the headline changed is offering a hypothesis that still needs checking. It cannot see a broken tag, a seasonal effect, or a change in your ad targeting unless you tell it, and it may fill gaps with plausible explanations. Ask it for several hypotheses and the evidence that would separate them. Check the underlying numbers yourself before acting, and test changes with a controlled comparison where you can.

c. A sensible workflow: Let AI propose and let data decide.

Use it to produce a list of candidate causes for a metric change, rank them by how cheaply they can be checked, and verify the top ones in your analytics and session recordings. Then run a test for the change you choose.

Where Episode fits

Episode is a platform for building and running campaigns to maximize conversion. It doesn't replace your analytics stack or your ad platforms, and it does not bid or target on your behalf. Where it helps with the landing page metrics to track above is in the build-measure-improve cycle for pages.

a. Building the page the metrics are about

Episode builds complete campaign pages and microsites from a campaign description and a company URL, and teams edit them visually without code. That shortens the time between spotting a problem in the data and changing the page, which matters when the data points to a page change, such as a different headline or form.

b. Conversion and funnel reporting on the page itself

Episode includes conversion and funnel analytics, heatmaps, and UTM and traffic-source reporting. These cover several behaviour and acquisition metrics in this guide, such as source-level conversion and where visitors lose interest, in the same place the page is edited.

c. Forms and lead capture

Forms capture leads and can send them to CRMs such as HubSpot, Salesforce, and Brevo. Connecting the form to your CRM is what makes qualified lead rate and cost per qualified lead possible to calculate by campaign.

d. Testing page and section variants

Episode supports experiments on pages and sections, so a hypothesis from the table above can be tested against the current version. Our article on how to improve landing page conversion rates lists the changes to try first.

e. Where it doesn't fit

Episode isn't an analytics system for your whole site, and it won't replace your CRM or ad platform reporting. If you rarely launch campaign pages, the gain is smaller. Revenue, pipeline, and ROAS will still come from your CRM and ad platforms.

Conclusion: Choose fewer landing page metrics to track and attach a decision to each

The landing page metrics to track are the ones tied to a goal, a decision, and a next step. Start with the outcome that matters for your page: conversions, qualified leads, revenue. Add a handful of diagnostics for the places visitors tend to drop off, keep the supporting numbers in the background, and read every metric by source and device before concluding that the page is the problem.

Before the next campaign goes live, define the primary conversion, tag every link, test the form on a phone, and decide how you will learn which leads were qualified. Then build the short dashboard, and review it on a schedule with the question that each number is meant to answer. Twenty numbers are available, and most pages need only a few of them in any given week.