You paid for a website. It came out looking good, you show it off proudly on WhatsApp, and yet the phone doesn't ring any more than it did before. The frustration is real and the cause is almost never a mystery: there are five numbers that explain it.
This post shows you what those numbers are, where to check them for free, and what to do when they come out bad. At the end I share the real metrics from Tekysoft's own site so you can see what we mean by good values.
The symptom: you have visitors but no clients
The typical case looks like this. Your site gets between 300 and 1,500 visits a month. You show up on Google when people search your name. People come in, look for three seconds, and leave. Nobody calls, nobody fills out the form, nobody books anything.
The owner usually assumes one of two things, both wrong. That they need more traffic, or that they need to redesign everything all over again. Most of the time it's neither. It's that the site fails on one or more of five measurable metrics.
The 5 metrics that matter (and why your site fails on them)
The first three are technical: they measure whether your site loads fast and responds well. The last two are behavioral: they measure what people do when they arrive. All five numbers matter equally.
LCP — how long your main content takes to appear
LCP (Largest Contentful Paint) measures the seconds from when someone lands until they see the largest element on the page: the main headline, the hero photo, the block that defines what the site is about.
Google's target: under 2.5 seconds. If your LCP sits at 4 or 5 seconds, one out of every two visitors leaves before the site finishes loading. That's not an opinion, it's what the aggregated Chrome data says.
Common causes: unoptimized 3 MB hero images, heavy fonts that block rendering, WordPress plugins that load scripts before the content, slow shared hosting.
INP — how long it takes to respond when they click
INP (Interaction to Next Paint) measures the delay between someone tapping a button or opening a menu and the moment the site reacts visually. It's the metric that replaced FID in 2024.
Target: under 200 milliseconds. If your menu takes half a second to open or the contact button feels sticky, people assume it's broken and leave. The human brain perceives anything over 250 ms as slow.
CLS — whether the layout jumps while it loads
CLS (Cumulative Layout Shift) measures how much the page elements move after they appear. The classic case: you try to tap a button, the cookie banner pops up on top, and you end up tapping something else.
Target: under 0.1. A high CLS isn't just annoying. It also lowers your Google ranking and drives up the exit rate because people feel the site is unstable. It's caused by images without declared dimensions, ads that get inserted late, and fonts that change size as they load.
Conversion rate of the main CTA
This one isn't technical anymore, it's business. What percentage of people who land do what you want them to? Book, call, fill out the form, request a quote. That action is your main CTA and every page should have a clear one.
Realistic ranges for a local business site in the US: between 2% and 5% is considered healthy. Below 1%, something is broken. If you get 1,000 visits a month and only 3 people fill out the form, it's not a traffic problem.
Common causes of low conversion: a CTA hidden below the fold, a form with 10 fields when 3 would do, a button with vague text like Submit instead of Book free, no WhatsApp button on mobile.
Bounce rate per page
Bounce rate is the percentage of visitors who land on a page and leave without touching anything else. Looked at across the whole site it doesn't tell you much. Looked at page by page, it tells you exactly where people drop off.
If the home page has a 40% bounce rate but the services page has 85%, you already know where the problem is. The services page probably loads slow, the copy doesn't answer the question the visitor came with, or there's no visible CTA.
How to measure them (free tools)
The three tools you need are free and official from Google. With these three you've got 90% of the diagnosis covered.
- Google PageSpeed Insights (pagespeed.web.dev). You enter the URL and it returns your real LCP, INP and CLS, measured in the browsers of actual users over the last 28 days. It measures mobile and desktop separately. Most US Hispanic traffic is mobile, so the Mobile tab is the one that matters most.
- Google Search Console (search.google.com/search-console). It shows you which searches you appear in, how many clicks you get, and the Core Web Vitals section with pages flagged as Good, Needs improvement or Poor. If you've never connected it, setting it up takes 10 minutes.
- Google Analytics 4 (analytics.google.com). It gives you conversions, bounce rate per page, and navigation paths. The important part is configuring your CTAs as Events: every click on Book, every form submission, every click on the WhatsApp number. Without defined events, GA4 tells you nothing useful.
What to do if the numbers are bad
Now you have the diagnosis. Here comes the part that moves the needle. These six changes are the ones that typically improve a site's total conversion by 20% to 60%. We don't promise exact figures because it depends on the business, but that order of magnitude is realistic.
- Optimize the hero images. Compress to WebP, set width and height in the HTML, and load the first image with priority. This single change usually cuts LCP by 1 to 2 seconds.
- Remove WordPress plugins you don't use. Every active plugin loads scripts. If you have 30 plugins installed and use 8, the other 22 are slowing you down without adding anything.
- Put a clear CTA above the fold on every page. One button, a contrasting color, text that states the benefit (Book 45 free minutes beats Contact us).
- Cut the form down to 3 fields: name, phone or email, and an optional message. Every extra field reduces submissions by 8% to 15%. You ask for the rest on the first call.
- Add a floating WhatsApp button on mobile. For US Hispanic businesses, WhatsApp converts better than any form. It's the customer's natural channel.
- Move from shared hosting to something decent. Hostinger Business is what we use on our projects: serious performance without overpaying. The difference between a $4 host and a $15 one shows up in LCP and in SEO.
Real case: how we measure the Tekysoft site
Practice what you preach. The site you're reading is built in Next.js 14 with server-side rendering, hosted on Hostinger Business with a CDN in front. These are the real metrics measured with PageSpeed Insights last week, Mobile tab.
- LCP: 1.8 seconds. Below the 2.5s threshold. The hero uses a 42 KB WebP image loaded with priority.
- INP: 140 milliseconds. The menu and buttons respond instantly because almost all the JavaScript is loaded lazily.
- CLS: 0.03. Practically zero. Every image has declared dimensions and the fonts load with font-display swap.
- Conversion rate of the Book 45 free minutes CTA: in progress, the first month of real traffic starts June 2026. We'll publish the real number once we have 60 days measured.
- Bounce rate per page: home 38%, methodology 45%, contact 22%. The methodology page is the one we'll optimize most in the next sprint.
These aren't showroom figures. They're the ones we have. We'll publish the conversion number when it exists, not before. That's the commitment: show what we measure, not what would sell nicely.
If your current site is far from these numbers, it doesn't mean it's broken beyond repair. It means there's clear room to improve, and the first step is to measure, not redesign.
