
When reports show more sessions while the CRM keeps receiving the same number of leads, marketing teams usually go back to the media plan: they change the audience, test a new creative, or shift budget. Before making any of those changes, it pays to check whether visitors landed on a page that was ready to use after the click.
Website performance helps answer that question. It shows how long the main content takes to appear, how the page responds to interactions, and whether elements stay in place while people use it. According to the HTTP Archive's 2025 Web Almanac, 48% of origins measured on mobile had good results on all three Core Web Vitals (Google's metrics for loading, responsiveness, and visual stability), based on July 2025 data. More than half fell outside the good range on at least one of them.
Technical performance explains only part of the result. Fast pages also convert poorly when the offer is unclear, the content doesn't answer the visitor's question, or the campaign attracts an audience with no interest in that service.
The sections below, with insights from Micaela L. Rossetti, Head of Marketing at Dexa, show how to separate these hypotheses, which technical and marketing metrics to read together, and how to decide what to fix first.
Getting visits but few leads?
Talk to Dexa about how your pages perform and what visitors experience on them.
Your site gets traffic, but what happens after the click?
Traffic is the first step in the marketing journey. When someone clicks your link, the expectation is that they'll keep moving until they become a customer, and often that doesn't happen. To find where potential customers drop off, you have to look closely at what people do after the click.
A session records a period of someone's interaction with the site. On its own, it doesn't tell you whether that person found the service they were looking for, understood the next step, or managed to submit the form.
On a B2B website, the path can go through three moments: entry, through a landing page, article, or homepage; research, on service pages, case studies, and pricing; and the contact attempt, through a form, chat, or meeting scheduler.
These three moments map to the table below: entry and research show up as visits, the contact attempt produces a lead, and sales decides which leads become opportunities.
Each funnel metric tells part of that story, and none of them tells all of it:
Metric | What it shows | What it doesn't show
|
|---|---|---|
Visit (session) | That someone started a session on the site | Whether the person was able to use the page |
Lead | That someone shared contact details through a conversion action | Whether the contact fits your buyer profile and is ready to buy |
Qualified opportunity | That sales recognized real business potential | How many people gave up before getting in touch |
When sessions grow and opportunities stay flat, the loss can sit at any handoff between these stages. Looking at each moment of the journey helps pinpoint where people struggle before you blame the media plan or the website.
What is website performance and why does it matter to marketing?
Website performance describes how pages behave while they load and while people use them. Google's Core Web Vitals assess three aspects of that experience.
Loading measures how long the largest content element visible in the initial viewport, such as the hero image or a block of text, takes to appear.
Responsiveness indicates how long the page takes to show visual feedback after a click, tap, or keyboard input.
Visual stability measures how much elements shift unexpectedly while the page is in use.

For marketing, these three dimensions sit between the ad and the conversion. Say an ad promises a free assessment, and the block that explains the assessment is slow to load on mobile.
During that gap, the visitor decides whether to keep staring at an empty space. When the ‘Request a proposal’ button takes almost a second to react, the person may tap again, assume the page froze, or leave.
Performance metrics don't confirm whether conversion flows work correctly. A form can display a success message without saving the lead to the CRM, a click-to-call or chat link can point to the wrong number, and a phone field can reject numbers that include a country code.
These functional failures call for a separate check, with manual tests that follow the path from the interaction on the page to the lead arriving in the CRM.
How can performance problems undermine campaigns and content?
In pay-per-click campaigns, the company has already paid for the visit by the time the person reaches the page. Problems from that point on can reduce the return on ad spend, including in brandformance strategies, which combine brand building and performance marketing and rely on the same page to carry both the brand message and the conversion.
Three problems show up often on landing pages:
Heavy above-the-fold content: The hero image or video delays the main content, especially on mobile, and the visitor never sees the value proposition.
Slow buttons: Ad pixels, chat widgets, and heatmap tools compete for the browser's processing power, and clicks get no visual response.
Shifting elements: A notice or a carousel pushes the form down just as the person taps, and the click lands in the wrong place.
Part of the loss never reaches the reports. When a page takes too long, the visitor leaves before the analytics tag fires: the ad platform charges for the click, but no session is recorded. A growing gap between clicks and sessions for the same campaign suggests the problem may be bigger than the numbers show.
Organic content works the same way. Articles with solid traffic and low conversion may be losing readers before the content appears, or because of a call to action (CTA) that moves around, and these cases rarely show up as errors in marketing reports.
Does website performance affect SEO?
Core Web Vitals are used by Google's ranking systems and are part of any search engine optimization (SEO) analysis. Good scores on these metrics, however, don't guarantee high rankings, as Google Search Central's documentation on page experience points out.
SEO analysis should combine performance, relevance, and content quality. For marketing, performance matters at two points: a slow page can lose visibility in search and, when it does get visits, make poorer use of the traffic it earned.
How can you tell whether low conversion is a website problem?
Before approving more budget or a new website project, leadership needs to know where the loss happens. Low conversion can come from the page, the offer, the audience the campaign attracts, or gaps in measurement, and each cause leads to a different decision: optimize the page, revisit the message, adjust targeting, or fix tracking.
To get there, the team cross-references three dimensions in its analytics tool: landing page, traffic source, and device. The table below shows how each combination of signals points to a hypothesis and what to check:
Signal | Hypotheses to investigate | How to check
|
|---|---|---|
Same page and same source convert well on desktop and poorly on mobile | Mobile experience, performance, or layout | Field data by device and a test on a mid-range phone |
One source converts poorly across all pages and devices | Mismatched audience or intent, or a misidentified source (UTM and attribution) | Search terms, targeting, and lead quality from that source, plus the UTM parameter setup |
All sources convert poorly on a page that's fast in field data | Offer, messaging, usability, or a broken flow | Submission test, page review, and offer or CTA testing |
Many form starts and few submissions | Long form, errors, or confusing validation | Error and abandonment events, plus a manual submission test |
Analytics records more conversions than the CRM receives | Measurement, duplicates, or integration | Comparison between the two databases, counting rules, and an end-to-end flow test |
Each signal needs context. A gap between mobile and desktop may reflect different audiences or campaigns, and a form with few submissions may be asking for information visitors aren't ready to share at that stage of the purchase. When the hypotheses point to the offer, messaging, or form, conversion rate optimization (CRO) tests help validate each change.
For decision-makers, the useful question to ask the team is which hypothesis was ruled out, and based on what evidence. A diagnosis that reaches a conclusion without these checks tends to recommend the solution someone already wanted.
“Often, what looks like a website problem is poorly targeted content or a tool that isn't being used well. That's why I start with behavior: with Hotjar and Google Analytics, I see where people drop off and validate the hypothesis before changing anything. In one assessment, only 2 out of every 100 visitors scrolled to the bottom of the homepage. The problem was engagement right at the top of the page,” explains Micaela L. Rossetti, Head of Marketing at Dexa.
Which metrics show whether your website is hurting conversion?
A useful report for leadership puts two kinds of information about the same page side by side: how it behaves technically and what visitors manage to do on it. Kept apart, these data points feed parallel discussions between marketing and development. Together, they show where the experience may be costing you leads.
On the technical side, the reference is the three Core Web Vitals. Google publishes thresholds for each one and rates a page as good when at least 75% of measured visits meet the threshold, with mobile and desktop assessed separately.
Metric | What it shows | Good | Needs improvement | Poor
|
|---|---|---|---|---|
LCP (Largest Contentful Paint) | How long the main content takes to appear | Up to 2.5 s | Over 2.5 s, up to 4 s | Over 4 s |
INP (Interaction to Next Paint) | How long the page takes to react to clicks, taps, and key presses | Up to 200 ms | Over 200 ms, up to 500 ms | Over 500 ms |
CLS (Cumulative Layout Shift) | How much content shifts without the user expecting it (unitless score) | Up to 0.1 | Over 0.1, up to 0.25 | Over 0.25 |
On the marketing side come the metrics for the same page: form starts and completions, submission errors, scroll depth to the CTA, and conversion rate by landing page. If INP gets worse on the same page template where form starts drop, the relationship deserves a closer look. The overlap doesn't prove cause and effect, but it shows where to test first.
Even your best-converting page may be losing leads. With poor LCP, the current rate works as a floor: people who gave up before seeing the content or before they could click never make it into the count.
Why isn't a PageSpeed Insights score enough to make decisions?
The score that usually gets passed around in meetings comes from Lighthouse, a test run under simulated conditions. It helps the technical team find problems and compare fixes, but it varies between runs and doesn't measure responsiveness to interactions, which depends on real use.
The closest picture of your customers' experience is the field data, shown in a separate section of the same report. It comes from the Chrome User Experience Report (CrUX), which aggregates real visits from Chrome users over the previous 28 days. This data has limits: it doesn't include other browsers, such as Safari on iPhone, and when a page has little traffic, it may show data for the whole origin (all pages on the same domain) in place of the URL you checked.
A lab score of 90 can coexist with poor field results, and the reverse also happens, as web.dev's guide to lab and field data explains. On landing pages for short campaigns, which rarely build up enough samples in CrUX, real user monitoring (RUM) tools can collect their own measurements. Whatever the number, the question for the team is which source it came from.
What should you fix first to get more out of your traffic?
Audits usually flag dozens of items, and fixing everything rarely fits the schedule or the budget. The order of fixes is a business decision, based on three criteria:
Page importance: Landing pages with active media, service pages, and the path to the form come before low-traffic corporate pages.
Severity: Functional failures that block conversion come first. Among pages of similar importance, “poor” results come before those that “need improvement”.
People affected: A mobile-only issue moves up the list when mobile accounts for most of the traffic or the media spend.
With that order in place, development estimates the effort for each fix, and low-effort fixes on important pages go first. Causes tend to cluster in four groups:
Images and video: Format, dimensions, compression, and the choice between video and a static image on mobile.
Third-party scripts: Pixels, chat, heatmaps, and A/B testing tools, often duplicated or left over from campaigns that have already ended.
Components: Carousels, pop-ups, embedded videos, and cookie banners that shift content or delay responses.
Integrations: CRM submission, scheduling, and validations that depend on external services.
Third-party scripts often require leadership involvement. Someone asked for each of those tools, and removing a pixel or a chat widget means negotiating with whoever depends on it.
“A heavy page, with large images and video, pushes away people who just arrived. When a campaign calls for video, we embed it from YouTube or Vimeo instead of uploading the file to the site, and images go in WebP. The campaign message is still there, without holding up the page load,” says Micaela.
How do you know whether a technical improvement paid off?
Whoever approves the investment in performance wants to know the return, and answering that takes care. Improving LCP and seeing conversion rise in the same month doesn't prove one caused the other: in the same period, media may have shifted audiences, the offer may have gotten a discount, and seasonality may have brought in buyers closer to a decision.
Three practices make the assessment more reliable:
Change log: Every code release, creative swap, offer change, and budget adjustment recorded with a date.
Like-for-like comparisons: Periods with the same mix of weekdays, no promotional dates on only one side, and audiences split by source, campaign, and device.
Controlled testing, when traffic allows: Visitors split between versions of the page, with other factors kept as similar as possible.
Timing matters too. Because CrUX uses a rolling 28-day window, field data takes about a month to fully reflect a fix. Expecting results the following week tends to lead to the wrong conclusion.
“When you change everything at once, there's no way to know what worked. I prefer to make one change, let the tools measure, and only then move on to the next. In larger strategies, such as a blog or paid media, an isolated test takes one to two months,” says Micaela.
Who owns website performance day to day?
Gains hold when the topic has an owner and set checkpoints. Without them, every new campaign, tool, or component can undo earlier improvements without anyone noticing.
Responsibility is shared. Marketing owns the tools, assets, and content it requests or adds. Development handles implementation, components, and infrastructure. Both teams keep the same change log.
Two moments concentrate the checks:
Before each campaign: Open the landing page on a mid-range phone over a cellular connection, send a test lead all the way to the CRM, confirm the tags fire without duplicates, and tap the main buttons, since lab tests don't measure responsiveness to interactions.
After significant releases, such as a new page template, a component, or a third-party tool: Run lab and submission tests right after the change and track field data over the following weeks.
When does a company need specialist help with website performance?
Many fixes are within reach of the in-house team: compressing images, removing tags from campaigns that have ended, testing the form before each launch. Outside help makes sense when the problem runs into limits of expertise, time, or platform. Some signs:
The diagnosis doesn't add up: The team has tested media and offer hypotheses, conversion is still low, and no one can say where the loss happens.
Marketing and development don't share data: Each team has its own reading of the problem, with no shared change log and no page-level metrics side by side.
The platform limits fixes: The CMS, theme, or architecture blocks basic adjustments, and every improvement turns into a project.
Media spend is about to grow: Scaling acquisition with a page that wastes clicks also scales the loss.
A redesign or migration is on the table: Performance, accessibility, and SEO are easier to secure when they're part of the project from the start.
In these cases, the gain comes from bringing acquisition data, page experience, and technical implementation into a single analysis, which is hard to do when each vendor handles a different piece.
How do you choose a partner to improve website performance and conversion?
Proposals focused on raising the PageSpeed score solve part of the problem. The choice gets safer when the partner shows, during the sales conversation, how it will connect performance, experience, and marketing results. Five criteria help compare proposals:
Diagnosis before solution: The partner asks for access to analytics, the CRM, and campaign data before recommending a redesign, a platform change, or new tools.
Multidisciplinary team: Strategy, UX, and development take part in the analysis, because the causes usually cut across all three.
Honest reading of results: The partner explains how it will measure the contribution of each improvement and doesn't promise Google rankings or a guaranteed lift in conversion.
Platform expertise: Experience with the CMS and integrations you use, or clarity about that technology's limits.
Defined continuity: The proposal makes clear who tracks the metrics after delivery and how new campaigns and tools will be evaluated.
In the meeting, three questions help test these criteria: What data do you need to get started? How will you tell a website problem apart from an offer problem? What happens to performance after the project ends?
How Dexa connects diagnosis, experience, and development
At Dexa, the strategic assessment analyzes journeys, experience, and conversion data to guide the recommendations. From there, the Experience Design team designs sites built for conversion and Core Web Vitals, and the Enterprise Technology team works on platform architecture and performance. The Digital Growth team tracks acquisition, content, and conversion to evaluate results and guide the next round of adjustments.
“The assessment starts from the client's goal: brand awareness calls for different improvements than lead generation. And not every client has the budget for a new site right now. So we propose phases: content and image adjustments in the short term, bigger changes in the medium term, and the new site in the long term,” explains Micaela.
Find out where the path from visit to lead is breaking down
Before you increase media spend, talk to Dexa's Digital Growth team to analyze performance, experience, and conversion and set the priorities for your site.
Frequently asked questions about website performance
What is website performance?
Website performance describes how pages behave while they load and while people use them. Core Web Vitals assess three aspects of that experience: loading of the main content, responsiveness to interactions, and visual stability. Functional tests complement the analysis by checking whether actions such as submitting a form complete correctly. Those tests aren't part of performance itself: they assess whether the flow works, not how fast the page is.
What's the difference between website speed and website performance?
Speed usually refers to load time. Website performance also covers aspects such as responsiveness to interactions and page stability. A site can load quickly and still be slow to react when someone clicks the contact button.
What are the thresholds for LCP, INP, and CLS?
A page is rated good with LCP of 2.5 seconds or less, INP of 200 milliseconds or less, and CLS of 0.1 or less. The assessment uses the 75th percentile of each metric, with mobile and desktop evaluated separately. CLS is a unitless score.
Does a PageSpeed Insights score show whether a site converts well?
No. The performance score comes from a lab test run by Lighthouse. The report also shows field data when there are enough samples, but none of these results measures conversion directly. Offer clarity, traffic quality, and form functionality require separate analysis.
Does improving website performance help Google rankings?
It can contribute. Google uses Core Web Vitals in its ranking systems, but good results on these metrics don't guarantee high rankings. SEO work should also look at relevance, content quality, and other technical factors.
When should you hire an agency to improve website performance?
When the in-house team can't find the cause of low conversion, when the platform limits fixes, or before scaling media spend, a redesign, or a migration. At those points, outside support helps bring marketing, experience, and development data into a single analysis.
How often should you monitor website performance?
A monthly review is a reasonable starting point, with extra checks before campaigns and after significant changes. Critical pages may need closer monitoring. This cadence is an editorial suggestion, not a Google rule. Functional failures should be checked right after a release, while field data needs time to reflect changes.






