Short answer
A page feels slow when the main photo or the headline takes too long to appear, when a tap does nothing for a moment, or when the page jumps while it loads. You can test that for free with PageSpeed Insights. It runs a lab test called Lighthouse, and it can show real Chrome visits from the last 28 days when enough people have used the page. This page is general information. It is not a guarantee that a faster page will rank higher or bring more calls.
Google says its core ranking systems look to reward content that provides a good page experience. Speed is one part of that experience. Core Web Vitals are three measures Google publishes for every page: how fast the main content loads, how the page responds when someone taps or clicks, and whether the page stays still. Google also asks whether the page is secure, whether it displays well on a phone, whether ads get in the way of the content, and whether a popup blocks the page. There is no single page experience signal. Google says Core Web Vitals are used by its ranking systems, and it recommends a good result both for Search and so the page is easier to use. A good score in a report does not put you first. Google says a strong result in Search Console, or in a testing tool, does not guarantee a top position. Search still tries to show the most relevant page even when the experience is poor. When many useful pages match the same search, a better experience can contribute. web.dev says a site owner should not have to be a performance expert to tell whether the page feels good. A plumber in Medina and a shop owner in Brunswick can read the three numbers below without learning a new trade.
Source: Understanding page experience (Google Search Central); Core Web Vitals and Google search results (Google Search Central); Web Vitals (web.dev)
Open PageSpeed Insights and paste the public address of the page, usually the homepage. Run the phone result and the desktop result. Google says the tool reports the experience on both, and that it suggests how the page may be improved. It shows two kinds of numbers. Lab data is one controlled test. It is useful for finding what to fix, and it can miss a problem that only shows up for a real visitor. Field data is what real people experienced. PageSpeed Insights takes field data from the Chrome User Experience Report. Google says that report is anonymized speed data from Chrome users who opted in, covering the previous 28 days when enough visits exist. The lab test is Lighthouse. web.dev says Lighthouse is available in Chrome DevTools, the tools that come with the Chrome browser. Google says PageSpeed Insights uses Lighthouse for the lab test, so that page runs the test for you and puts the real visit numbers beside it. For a first look, PageSpeed Insights is the simpler free test. For the phone lab test, Google says Lighthouse pretends to be a mid-tier phone, which the docs name as a Moto G4, on a mobile network. The desktop lab test pretends to be a desktop computer on a wired connection. The test runs in a Google data center, and the report names the region as North America, Europe, or Asia. Paste the public address. A landscaper, a plumber, or a shop can use the same page.
Source: PageSpeed Insights; About PageSpeed Insights (Google); Getting started with measuring Web Vitals (web.dev); Web Vitals (web.dev)
At the top of the lab section, PageSpeed Insights shows a performance score from Lighthouse. Google says 90 or above is good, 50 to 89 needs improvement, and below 50 is poor. A good lab score does not mean real visitors had a good experience. The score can also change from one run to the next when you did not change the page. Google says the network and the computer running the test both vary. If the number jumps, run it again and read the suggestions, not only the score. The real visit section uses the 75th percentile. That is a slower visit, chosen so the page is judged by more than a fast laptop at the shop. web.dev says that for a page to meet the Core Web Vitals recommendation, 75% of page visits should be in the good range for each of the three measures, and that phones and desktops are counted separately. PageSpeed Insights says the page passes when the 75th percentile is good for all three. A small or new site often has no real visit numbers. Google says the page needs enough Chrome users in that 28 day window. If the page is too thin, PageSpeed Insights tries the whole site. If the whole site is also too thin, it shows no real user data. That is a limit of the sample. The lab report and the suggestions are still there to use. Lighthouse reports Largest Contentful Paint and Cumulative Layout Shift in the lab. It does not measure Interaction to Next Paint there, because that measure needs a real tap. web.dev says the lab stand-in is Total Blocking Time, the time during the load when the page is too busy to respond. Open the page on a phone you own as well. The score is a speed test. The phone guide on this site is the check for whether the words fit the screen.
Source: About PageSpeed Insights (Google); Web Vitals (web.dev); Getting started with measuring Web Vitals (web.dev)
On this site: How to check if your website works on phones
Largest Contentful Paint, shortened to LCP, is the loading measure. web.dev says it is the time from the moment the page starts loading until the largest image, or the largest block of text, in the first screen is shown. A good LCP is within 2.5 seconds. Interaction to Next Paint, shortened to INP, is the response measure. It watches taps, clicks, and key presses during the visit. The number it keeps is a long interaction from that visit, with rare outliers left out. A good INP is 200 milliseconds or less. A millisecond is one thousandth of a second, so 200 milliseconds is 0.2 seconds. web.dev says a poor INP is greater than 500 milliseconds, and the band between is labeled needs improvement. Cumulative Layout Shift, shortened to CLS, is the stillness measure. It scores how much of the visible page moved, and how far. It is not a clock. A good CLS is 0.1 or less. web.dev says a poor CLS is greater than 0.25. PageSpeed Insights uses those same good cutoffs, and it calls LCP poor when it is over 4 seconds. The middle band is Needs Improvement. Judge phones and desktops separately, at the 75th percentile, so a fast office computer does not hide a slow phone.
Source: Web Vitals (web.dev); Optimize Largest Contentful Paint (web.dev); Optimize Interaction to Next Paint (web.dev); Optimize Cumulative Layout Shift (web.dev); About PageSpeed Insights (Google)
A huge photo is a common reason a small business homepage waits. WordPress.org says the file size of an image sets how long the page takes to load. A larger file, often from a high resolution photo, takes longer, and people often do not wait through a long load. The same article says JPEG is a good fit for a photograph, because saving as JPEG removes some detail and a photo editor lets you choose how much. PNG keeps that detail and usually makes a larger photo file than JPEG. The WordPress optimization handbook says to pick the format and the compression for each image, to ask whether a picture could be a sentence instead, and to consider WebP, which is smaller. The image article says large high quality images should typically stay between 100K and 60K, and smaller images closer to 30K and lower. That is their published target for a file you upload. web.dev says compressing an image, or using a smaller format such as WebP or AVIF, shortens the download. AVIF is another newer image format web.dev groups with WebP. The largest image in the first screen is often the LCP element, the one the loading score is waiting on. Since WordPress 5.5, images wait until they are about to scroll into view. WordPress.org calls that lazy loading, and says that on a phone it can also skip files meant for a different screen. Leave that wait off the big photo at the top. web.dev says lazy loading on that image holds the download until the browser has confirmed the photo is in the first screen, so the largest content starts later.
Source: Image size and quality (WordPress.org); Optimization (WordPress.org); WordPress 5.5, lazy-loaded images (WordPress.org); Optimize Largest Contentful Paint (web.dev)
A plugin is an add-on you installed in WordPress. WordPress.org says the number of plugins, and how well they perform, has a huge impact, and that turning off and deleting plugins you do not need is a significant way to speed the site up. Turn them off one at a time and run PageSpeed Insights again, so you can see which one changed the result. If one plugin is the problem, WordPress.org says to read its documentation, ask in its support forum, or find another plugin that does the same job. A theme is the design package. WordPress.org says a fast lightweight theme performs much more efficiently than a heavy, graphic-laden one. A page builder is a plugin or a theme that lets you drag sections onto the page. WordPress.org does not publish a separate rule for builders. It does say that a heavy theme and extra plugins both cost speed, and that you should cut the number of files on a typical page and minify CSS and JavaScript. CSS is the file that sets colors and layout. JavaScript is the file that makes something move or respond, such as a menu. Minify means making those files smaller by removing space the browser does not need. Large JavaScript also delays the biggest photo. web.dev says the browser has to read and run that JavaScript on the main thread, the single sequence of work the page can do at a time. An image can finish downloading and still wait until an unrelated script is done before it appears.
Source: Optimization (WordPress.org); Optimize Largest Contentful Paint (web.dev); PageSpeed Insights
Shared hosting means your site sits on a server with many other sites, and the host manages that server. WordPress.org says this is the most common setup, and that you have very little control over the server settings. A low cost shared host still needs an efficient site. Hardware matters. WordPress.org says hosts generally offer higher performance for a higher price, and that if you are on shared hosting and you are maxing out the plan limits, a plan with higher limits can help. The distance between the server and your visitors also changes the wait. Time to first byte is how long the server takes to start answering. WordPress.org uses that name for the delay from extra database trips. A database is where WordPress stores the pages and the settings. web.dev says a high time to first byte adds time to the loading measures after it, including the largest content. It also says 0.8 seconds or less is a rough guide, and that time to first byte is not one of the three Core Web Vitals. Caching means keeping a saved copy so the same work is not repeated. WordPress.org says a caching plugin saves posts and pages as static files, a finished copy the server can hand out, and that this reduces the processing load. It also says that without a caching solution, performance slows as more requests come in and stack up. Browser caching is the copy on the visitor computer. If images, CSS, and JavaScript send the right cache headers, a note that says the browser may keep the file, the browser can reuse them. The server can then say the file has not changed, instead of sending the whole file again. A persistent object cache saves repeated database lookups. WordPress.org says that without one, the server reads options from the database on every view, which slows time to first byte, and that the host must offer a cache server before that cache can be installed. Ask the host whether a cache is already on, and whether the account is hitting its limits. Changing the server is their job, or the job of someone who manages the site. WordPress.org says that work is often outside what a hosting plan includes.
Source: Optimization (WordPress.org); Optimize Time to First Byte (web.dev)
web.dev says the browser starts downloading a video when it finds the video tag on the page. Autoplay is the setting used when a video must play immediately, such as a clip running behind the words as a background. An animated GIF, a moving image in an older format, can be very large. web.dev says it is not uncommon for one to use several megabytes, and that a video file is usually a more efficient way to show motion when you need motion at all. A slider, which web.dev calls a carousel, often includes photos that are not on the first screen when the page opens. Those photos can still be requested early. web.dev says to give the off screen slides a lower priority so they do not compete with the main image. Photos and video also make the page jump when they have no width and height. web.dev lists images without dimensions as a common cause of layout shift, and says to set width and height so the browser can reserve the space. A YouTube or Vimeo box is an embed from another company. web.dev says those players often load a lot of extra JavaScript, and that a long block of that work can hurt Interaction to Next Paint. If the video does not need to play by itself, turn autoplay off, or take the video off the first screen. If the slider is only decoration, take it off and leave one photo of a real job.
Source: Video performance (web.dev); Optimize Largest Contentful Paint (web.dev); Optimize Cumulative Layout Shift (web.dev)
web.dev says a third party script is code you did not write, served from someone else's server. Chat services, social sharing buttons, video players, and ad frames are the examples it names. A review badge that loads from another company is that same kind of script. Any significant amount of JavaScript can slow a page. These scripts add extra trouble because you do not control them. They open more connections. That setup takes time, and a secure connection can take longer. They can pull in images and video that were not compressed, and they are often cached poorly, so the browser has to fetch them again. If the script is loaded in a way that makes the browser wait, it delays reading the rest of the page. If that other server fails to answer, web.dev says the page can stay blocked until the request gives up, which can be anywhere from 10 to 80 seconds. Pick the option that sends the least code for the job you actually need. web.dev says you probably do not need two products that do the same job. Its example is two tag managers, tools that load other tracking scripts, or two analytics tools, tools that count visits. Two chat boxes, or two review badges, are that same kind of overlap. These boxes can also shove the page around. web.dev says a YouTube video, a map, or a social post often does not know how big it will be before it loads, so the content under it jumps when the box appears. Reserve space for a chat box or a badge, or leave it off until you know you want it.
Source: Third-party JavaScript performance (web.dev); Optimize Cumulative Layout Shift (web.dev)
You can do a useful pass yourself when you can log in. Save photos smaller before you upload them. Use JPEG or WebP for photographs, and aim at the file sizes WordPress.org publishes. Leave the top photo so it loads right away. If the editor has a switch for lazy load on that photo, turn the switch off. If you cannot find the switch, leave that one edit to the person who looks after the site. Turn off plugins you do not need, one at a time, and run PageSpeed Insights after each change. Remove an autoplay video, a slider, or a chat box you added, when the page still names the work, the towns, and the phone number. Ask the host, in writing, whether caching is on and whether the account is hitting its limits. Hand the rest to someone who does this work. On shared hosting you have very little control over the server settings. A persistent object cache needs a cache server from the host. Minifying files, replacing a heavy theme, and marking which image should load first are edits for the person who looks after the site. Retest after each change they make. One change at a time is how you learn what helped.
Source: Image size and quality (WordPress.org); Optimization (WordPress.org); Optimize Largest Contentful Paint (web.dev); PageSpeed Insights
Read this on the site you have now. You ran PageSpeed Insights for the phone and for the desktop, and you wrote down the performance score and LCP, INP, and CLS. If the real visit numbers were missing, you noted that and used the lab suggestions. The main photo is a reasonable file size, in JPEG or WebP, and it is not set to lazy load. Plugins you do not use are off. Video on the first screen does not play by itself. A slider, if you still need one, is not a stack of huge photos. You have one chat or review tool, or none, and not two that do the same job. You have asked the host whether caching is on. You opened the page on a phone you own and could find the phone number without a long wait. If a line fails, that line is the next job.
Source: PageSpeed Insights
On this site: How to check if your website works on phones
The free website checker reads one public page and reports HTTPS, response status and time, the title, and a mobile viewport tag. A viewport tag is the line that tells the browser to use the screen width. The checker logs the website address and the score, not personal data. It leaves LCP, INP, and CLS to PageSpeed Insights. Website Quick Diagnostic is $19, paid once. An automated PDF within about 1 hour with your speed and mobile scores, the top 5 issues ranked, and a plain fix for each. The $19 is credited toward the $99 Checkup within 30 days. Website Checkup is $99, paid once: a ranked report of what to fix on your website first, covering speed, SEO, mobile, and AI search visibility, delivered within 2 business days. No login is needed. The report goes only to you. If the site is down and there is nothing to check, you get a refund. Those checks describe a page they can fetch. They do not resize your photos or change the host. If you want a new site, ask on the start page.
Source: PageSpeed Insights
On this site: Free website checker; Website Quick Diagnostic; Website Checkup; Start a site quote
The free website checker reads one public URL. It reports HTTPS, a mobile viewport tag, and whether GPTBot, ClaudeBot, PerplexityBot, and Google-Extended are blocked. It logs the website address and the score, not personal data.
The AI search readiness check is the free look for AI search. It reports whether OAI-SearchBot, ChatGPT-User, and PerplexityBot are blocked, and whether the words are in the server HTML. It logs the website address and the score, not personal data.
Website Checkup is $99, paid once. A ranked report of what to fix on your website first, covering speed, SEO, mobile, and AI search visibility, delivered within 2 business days. No login is needed. The report goes only to you.
Google says Core Web Vitals are used by its ranking systems, and that a good page experience can contribute when many pages are relevant. Google also says there is no single page experience signal, that Search still tries to show the most relevant page when the experience is poor, and that a good score does not guarantee a top position. This page does not promise a ranking or a set number of calls.
For the Lighthouse lab score, Google says 90 or above is good, 50 to 89 needs improvement, and below 50 is poor. A good lab score does not mean real visitors had a good experience. For the three Core Web Vitals, a good LCP is within 2.5 seconds, a good INP is 200 milliseconds or less, and a good CLS is 0.1 or less.
No. Google says a page needs enough Chrome visits in the previous 28 days before PageSpeed Insights can show real user data. A new or quiet page often does not have that. The tool then tries the whole site, and if that is also too thin it shows no real user numbers. Use the lab report and the suggestions.
Often, yes, for the photos, a plugin you do not need, a slider, an autoplay video, and a chat box you added, if you can log in. Server settings, a cache you are not sure how to install, and a heavy theme are jobs for the host or for someone who edits the site. Retest after each change.
No. The free checker reports response time for one public page, plus HTTPS and a few public tags. It logs the website address and the score, not personal data. It leaves LCP, INP, and CLS to PageSpeed Insights. Website Quick Diagnostic is $19, paid once. An automated PDF within about 1 hour with your speed and mobile scores, the top 5 issues ranked, and a plain fix for each. The $19 is credited toward the $99 Checkup within 30 days. Website Checkup is $99, paid once: a ranked report of what to fix on your website first, covering speed, SEO, mobile, and AI search visibility, delivered within 2 business days.
No. A page that appears quickly, responds to a tap, and stays still is easier to use on a phone. It does not promise a ranking or a set number of calls.
A business can be missing from Google because the page is not in the index, because the words you searched are not the words a customer uses, or because the page is blocked. A page can also be in the index and still sit far down the results. The free check on this site reads the public page, and it does not look up a rank.
You can check a public page for free. The checker on this site reads one URL, reports what it found, and logs the website address and the score, not personal data. If you want a report that ranks what to fix first, Website Checkup is $99, paid once, and it is delivered within 2 business days.
If you build it yourself, published US prices checked for this page start at $4 a month on WordPress.com when billed annually, $17.77 a month on Wix Light for a yearly plan, or $19 a month on Squarespace Basic billed annually ($228 for the year). If you hire an agency, Clutch says web design agencies typically charge $2,000 to $100,000, and that most reviewed projects cost less than $10,000. If Telep IO builds it, the founding offer is $500 to set up, then $50 a month. Limited to the first 3 Ohio businesses. The start page splits that setup into two payments and makes the first upkeep month free, which comes to $1,050 in the first year, then $50 a month.
Website maintenance is the work that keeps a live site reachable and current: hosting, domain renewal, SSL, backups, software and plugin updates, security, small content edits, uptime checks, and keeping a Google Business Profile matched to the site. You do not always need a monthly plan. You need one when you will not do that work, or when a fee is the only way the site stays hosted and edited. Other providers set their own prices. This page does not print those prices. The only prices stated here are the ones Telep IO publishes.
Before you hire, ask who owns the domain and the site, what hosting costs, who edits it later, whether the page works on a phone, whether they will help with a Google Business Profile, and when you will see a first look. Walk away if you will not own the domain, if the contract locks you in, if there is no portfolio you can open, or if the monthly fee does not say what it covers. Other people set their own prices. Ranges vary. The only build price stated here is the one Telep IO publishes.
Start with the domain. The registrar is the company that holds the registration. The registrant is the person or business the record lists as the holder. ICANN's free lookup shows the registrar and the contact data that registrar publishes. A hidden name, or a privacy service, is not proof that you own the name. If a former designer, a relative, or an agency stopped answering, write down the lookup, then ask that registrar about the account or a transfer. If the designer refuses, talk to a lawyer. This page is not legal advice.
First check whether the site fails for everyone, or only on your phone and your network. Then match the words on the screen to the usual cause, and contact the person who can fix that cause. This page is general information. It is not a guarantee that a step will bring the site back.
Redo the site when it no longer matches the business, you cannot edit it, or a customer on a phone cannot read it and call. Many of the usual problems can be fixed without a full rebuild. Old hours, a missing call link, a page that is hard to read on a phone, and an address that is not on HTTPS are often edits. A new look does not, by itself, put the business on Google or on Maps.
A visitor can miss you when the phone number is wrong, the hours are wrong, the form does not send, the page does not fit a phone, or the browser warns about the connection. An old year in the footer is a reminder to check those facts. Google says content is rated by quality, regardless of age, and that a trivial change to look fresh has no additional value. This page is general information. It is not a guarantee that fixing the page will rank higher or bring more calls.
A DIY builder can be the right call when you will finish the page, you will own the domain, and a customer on a phone can tap to call, see the towns you serve, and open your Google Business Profile. Hire when you will not keep the page current, when the number cannot be tapped, or when the domain and the site would sit in an account you do not control. Wix, Squarespace, and the GoDaddy website builder are typical monthly subscription plans. This page does not print their prices. Other designers set their own prices. The founding offer is $500 to set up and $50 a month, limited to the first 3 Ohio businesses. Hosting and ordinary upkeep are included. You see a demo of your own site before you pay.
Open the site on a phone you already own and try to call the business from the page. If you have to pinch and zoom, or you cannot find the number, the page is hard to use on a phone. A viewport tag tells the browser to use the screen width. It is a setup signal, not a grade of the layout.
Send one test from a phone, then look in the inbox, in spam, and in the list of messages the form saved. A missing lead can be the message sitting in spam, going to an old address, or never leaving the page. This page is general information. It is not a guarantee that every message will arrive, or that a fix will bring more calls.
A menu item that opens a page not found screen does not show the jobs, the towns, or the phone number on that page. A customer can leave before they call. MDN calls that screen a 404 when the server cannot find the page. Google says a 404 will not, in general, hurt search performance when the address should not exist. A link you still show in the menu is worth fixing because a person taps it. This page is general information. It is not a guarantee that fixing the link will rank higher or bring more calls.
A customer who opens the address and sees coming soon, under construction, a parked page, or a builder message that the site was not found does not see the jobs, the towns, or a phone number to tap. Someone who found the business on a Google listing, or who heard the name from a neighbor, can leave before they call. Google says a success code does not guarantee indexing. A placeholder with little content about the business can look like a page Google excludes from Search when Google treats it as a soft 404. A private or unpublished builder site can also stay out of search results. The builder help page says which setting does that. This page is general information. It is not a guarantee that publishing the real page will rank higher or bring more calls.
A plumber, a roofer, a landscaper, or a contractor can open the site and find pages about gambling, pills, or a language the business does not use. Google results can show a spam title, or the words "This site may be hacked." The browser can say "Deceptive site ahead." The address can send people to another site. WordPress can show "There has been a critical error on this website." A warning scares the person who was about to call. Google says hacked content is placed on a site without permission, and that it tries to keep that content out of search results. Google also says sites that violate its spam policies may rank lower or not appear in results at all. Hacked content is one of those policies. This page is general information. It is not a guarantee that cleaning the site will remove a warning, rank higher, or bring more calls. Treat it as general guidance. If you are not sure which files are the real site, hire help or ask the host.
An AI tool can mention your business when it has public information to use, or when a person typed the facts in. A robots.txt file can ask specific crawlers to stay off the site. That file is a request, not a record of what a tool has already saved. Blocking one crawler does not block every AI product.
A Facebook page is not a website you control. Google's guidelines say the website on a Business Profile should represent the business, and they say not to use a page created on a social media site as that website. A site you control can name the work, the towns, and a phone number a person can tap.
A local site should say what you do, name the towns you actually serve, and put a phone number a person can tap. It should also load on a phone. That is not a promise of a ranking or a set number of calls.
Map results come from a Google Business Profile, which is separate from the website. Google says that when you add and verify a Business Profile, customers can find the business on Search and Maps, and that only verified businesses can show their info there. A website does not create the profile. This page does not create it for you.
Ask a real customer after the job is done, and give them a direct review link from your Google Business Profile. Google lets you share that link, or a QR code, in a text or email, on a receipt, and at the counter. Do not pay for a review, do not offer a discount for one, and do not ask only the people who already said they are happy.
A website does not promise a set number of calls. For an HVAC company, a plumber, or a landscaper, the page can make the call easy: the number is tappable, the jobs are named, and the towns are the ones you drive. If those facts are buried, the person on the phone has to hunt.
Know an owner who needs this? Share it on X, Facebook, or LinkedIn.
A monthly email of short practical website tips for Ohio small businesses and trades. One email a month. Unsubscribe anytime.