Imagine this example: you run your flower shop website through Google's speed test on a Monday and it scores 42 on mobile. You run it again on Tuesday and it scores 57. Nothing on the site changed. A friend's web designer sends you a different tool's report that says the site is fine.
All three can be telling the truth. They are measuring different things in different ways. This guide explains what each speed number means, why they move about, and which ones are worth your time when a customer is ordering on a phone.
The Three Kinds Of Speed Number
1. Lab Scores (Google's Speed Test)
Google’s speed test loads your page once, on Google's network, pretending to be a mid-range phone on a slow 4G connection. It then gives you a performance score out of 100. Google colours 90 and above green ("good"), 50 to 89 orange ("needs improvement") and under 50 red ("poor").
That score is a blend of several timings from the same single visit, including:
- Largest Contentful Paint (LCP): how long the biggest thing on the first screen takes to appear. On a florist's homepage this is usually the banner photo.
- Total Blocking Time (TBT): how long scripts keep the phone too busy to respond to a tap while the page loads.
- Cumulative Layout Shift (CLS): how much the page jumps around as it loads.
- First Contentful Paint and Speed Index: how soon anything appears, and how quickly the screen fills in.
A lab test is useful because it tells you what to fix. Its weakness is that it records a single visit from one simulated phone, so the next run can come out differently.
2. Real-Visitor Data (The Chrome UX Report)
On the same results page from Google's speed test, look for the section about what your real users are experiencing. This comes from the Chrome UX Report: timings Google collects from Chrome users who have chosen to share their browsing statistics with Google, gathered over the last 28 days.
Google reports these figures for 3 in 4 visits (the 75th percentile). So "LCP 2.4 seconds" here means three quarters of real visits saw the main picture or text within 2.4 seconds. The three figures that make up Google's Core Web Vitals, with Google's "good" and "poor" lines, are:
| Measure | What it means | Good | Poor |
|---|---|---|---|
| Largest Contentful Paint | Main picture or text appears | 2.5 seconds or less | over 4 seconds |
| Interaction to Next Paint | Screen answers a tap | 200 milliseconds or less | over 500 milliseconds |
| Cumulative Layout Shift | Page holds still as it loads | 0.1 or less | over 0.25 |
Real-visitor data is the nearest you will get to what your customers see. The catch for most independent florists is that Google needs enough Chrome visits before it shows anything. Many small shop sites see "no data" for the homepage, or only a figure for the whole site rather than that page. That is normal and says nothing bad about your shop.
3. A Full Website Check
A full check loads your site in a real browser and measures more than speed. Our free Florist Website Checker is one of these. It runs Google's speed test on your homepage, on a phone and a computer, and adds its own visit in a browser on a phone-sized and a laptop-sized screen. How it weighs those results is covered further down.
Why The Numbers Differ Between Tools And Runs
If your score moves by several points between two runs and you haven't changed anything, that is usually normal variation. A lab score is one visit, and several things change from one visit to the next:
- How quickly your server answers. If the page was visited recently, your host may hand back a ready-made copy. If not, it builds the page from scratch, which takes longer.
- Other companies' code. A review widget, a live chat box or an embedded map loads from someone else's server. If that server is slow for a moment, your score drops.
- Different pretend devices. One tool might test as a slow phone on 4G, another as a laptop on fast broadband. As an illustration, the same page might score 45 on one and 90 on the other.
- What each tool counts. Some time how long until the page looks ready, others until every file has downloaded.
Real-visitor figures move slowly because they cover 28 days of visits. That has a knock-on effect after you fix something: a lab score can improve the same day, while the real-visitor figures can take up to a month to catch up.
A Worked Example
Here is an illustrative example for a fictional shop, not real measurements. Example Street Flowers runs Google’s speed test three times one morning:
| Run | Mobile score | Lab LCP |
|---|---|---|
| 9:05am | 41 | 6.8 seconds |
| 9:20am | 55 | 5.1 seconds |
| 9:40am | 48 | 5.9 seconds |
Its real-visitor section shows an LCP of 3.6 seconds for the site as a whole, rated "needs improvement".
The scores disagree, but every run says the banner photo arrives too slowly. The real-visitor figure may be quicker because visitors use a different mix of devices, connections and pages from the lab test. It is still over Google's 2.5-second "good" line, so that is the number to work on, and the lab runs show why.
Which Numbers Matter For A Flower Shop
For a florist, speed matters most at the moments a customer is in a hurry: a same-day birthday order on the bus, or the evening before Mother's Day. Many of those orders start on a phone, and your own website or shop figures can show you how many of yours do.
So, in rough order of importance:
- Real-visitor figures, if Google has them. These describe eligible Chrome visits to your site, which may include people who never order. Look at the phone figures first.
- Largest Contentful Paint. How long before a customer sees your flowers. On a photo-led site this is usually the number with the most room to improve.
- Layout shift. A page that jumps as it loads makes people tap the wrong thing, such as the wrong bouquet size or a banner instead of Add to basket.
- Tap response. Interaction to Next Paint from real visitors, or Total Blocking Time in the lab. A phone that freezes for a second after every tap feels broken.
- The overall performance score. Useful as a rough before-and-after. It is made from the numbers above, so work on those and the score follows.
You don't need 100. Google's own bands put 50 to 89 at "needs improvement" and under 50 at "poor", so lift a red score into orange first, then work on the real-visitor figures.
What Typically Slows Florist Websites
Florist sites rely on photos, because customers choose bouquets by looking at them. The problem is usually the size and format of each photo.
Large Photos Straight Off The Phone
A modern phone camera produces photos several thousand pixels wide. Uploaded as they are, a single banner can weigh several megabytes, which a customer on a phone downloads before your homepage looks ready.
As an illustration: a 4,000-pixel-wide banner photo at 5 MB, shown in a space about 1,400 pixels wide on a laptop and 400 pixels wide on a phone, could usually be exported at about twice the width of the space it fills (around 2,800 pixels here, to allow for sharp laptop screens) and a few hundred kilobytes, and look the same to most customers. Keep the full-size original for print and social media, and upload a web-sized copy.
Other photo habits that add weight:
- Old file formats. WebP and AVIF photos are usually much smaller than JPEG or PNG for the same picture. Many website builders convert them for you, but not all.
- Every photo loading at once. Photos far down the page can wait until the customer scrolls to them (called lazy loading). If all 40 product photos on the homepage download straight away, the banner competes with them.
- Photos with no set size. If the page doesn't know how big a photo will be, it moves everything when the photo arrives. That is where a lot of layout shift comes from.
Our flower photography tips for florists cover taking the photo in the first place.
Pop-Ups On The First Screen
A sign-up pop-up, a discount box or a large cookie notice that covers the phone screen as the homepage opens means a customer sees no flowers until they deal with it. It often brings extra code with it too. If you use a sign-up or discount pop-up, keep it small, or wait until the customer has had a look round.
Whether your cookie notice meets the rules is a separate question from how much of the screen it covers. Check that one with whoever looks after your website, or a data protection adviser.
Apps, Widgets And Embeds
Each extra service you add brings its own code from another company's server. On florist sites the usual ones are review widgets, live chat boxes, social media feeds, embedded maps and old tracking tags nobody looks at any more. Each one adds weight, and several together can outweigh the rest of the page and keep the phone busy, which slows tap response. Once a year, remove anything you no longer use. A map is often better as a link to your Google listing on the contact page than as a live map on the homepage.
A Slow Server Response
Before a single photo loads, your web host has to start sending the page. If that takes the best part of a second, everything else waits. This is usually down to hosting or setup rather than anything you upload, so raise it with your website provider or web designer.
How The Florist Website Checker Reads Speed
The free website check for florists starts with a quick check of your homepage. After you enter your web address and the check finishes, you see an overall score and a score for each of the seven areas it could check. Speed sits in Speed & Mobile.
Confirm the site is yours, for example with your shop email address, and you get the full report: every finding with how to fix it, marked on screenshots of your own site where a problem can be shown. For speed, that includes a strip of frames from Google's phone test, so you can see what a customer is looking at while the page loads.
Speed results come from Google's data where it has it, and from our own visit when it doesn't. Each result says where its figure came from.
Google's overall mobile score is shown as a headline but is not added to your score a second time, because the timings it is made from are already scored on their own. The same goes for Google's overall real-visitor rating.
The checker also measures the causes covered above: total photo weight on the homepage, photos far bigger than the space they fill, old photo formats, photos that load before they are needed, weight from other companies' services, slow server response and, under Design & Content, pop-ups covering the phone homepage as it opens.
The full report also includes an AI review that reads your pages and screenshots. It is clearly labelled as a second opinion, counts for less than measured results and does not judge speed.
Tips For Testing Your Own Site
- Test the phone version first. Google’s speed test opens on the mobile result, and it is the version a customer ordering on a phone sees.
- Run a lab test three times and look at the middle result, not the best or worst one.
- Check real-visitor figures a month after a fix, not the next day.
- Look before peak. Make speed changes a few weeks before Valentine's Day or Mother's Day, not during the week itself.
Common Questions
Why does Google’s speed test say there is no real-user data for my site? Google only shows real-visitor figures once enough Chrome users have visited in the last 28 days. Many independent florist sites don't reach that level for the homepage. You can still use the lab test to find what to fix.
Will a faster site get me more orders? A faster site gives customers less reason to give up on a phone. How many extra orders that brings depends on your shop and the rest of the site, so compare your own orders before and after.
Does speed affect where I appear on Google? Google says its ranking systems use Core Web Vitals, alongside many other factors. Good scores do not guarantee high rankings. For local searches like "florist near me", Google says results mainly follow relevance, distance and prominence, and reviews count towards prominence. Our guide to getting more Google reviews for your flower shop covers that side.
Can I fix speed myself, or do I need a web designer? Photo sizes, unused widgets and pop-ups are often things you can change yourself in your website's settings. Server response and how the site's code loads usually need your website provider or a web designer. The full report from the Florist Website Checker gathers the settings only your website provider can change in a section of their own, headed Things Your Website Provider Can Help With, so you can forward it to them in one go.
