What a slow site is costing you today
A site loading in 4.2 seconds versus a competitor loading in 1.8 seconds, same product, same price. On a 100-visitor day, that gap alone costs you 12 to 15 customers who leave before the page even finishes appearing.
That is not an exaggeration. Google's research on mobile page speed shows a one-second delay typically costs about 7 percent of sales. At 4.2 seconds, you are losing roughly 22 percent of the business that was already on its way to you.
This guide covers what that costs in real numbers, a 10-point checklist you can hand to whoever maintains your site, and the Cuez case where I cut response time from 3 seconds to 300 milliseconds. By the end, you will know exactly what to check and what it is worth fixing first.
TL;DR
Every 1-second delay costs you about 7 percent of sales. Google measures speed with three numbers: how fast your main content shows up, how fast the site responds to a click, and whether things jump around while loading. The biggest wins are caching, smaller images, loading things only when needed, deferring scripts you do not need right away, a global content network, and cutting scripts you no longer use. Fixing this usually costs $2,000 to $15,000 and pays for itself in 6 to 12 months through recovered sales. The Cuez fix made a product 10x faster (3 seconds to 300ms) and cut infrastructure cost by about 40 percent.
What a slow site actually costs you, in dollars
Speed is a revenue line, not a nice-to-have.
Sales lost to delay
| Delay | Sales lost | On 100 visitors, that is |
|---|---|---|
| 1 second | 7% | 7 customers |
| 2 seconds | 14% | 14 customers |
| 3 seconds | 21% | 21 customers |
| 4 seconds | 28% | 28 customers |
| 5 seconds | 35% | 35 customers |
A worked example. A store averaging $500 a sale, 100 visitors a day, a 10 percent baseline conversion, so 10 customers and $5,000 a day.
If the site loads in 4 seconds instead of 1, that costs 2 customers a day, or $1,000 a day, $30,000 a month, $365,000 a year. Fixing the speed to under 2 seconds typically costs $5,000 to $10,000 and pays back in 10 to 20 days.
What slow customers actually do
About 53 percent of mobile visitors abandon a site that takes longer than 3 seconds to load. Pages loading under a second convert at roughly 2.5 times the rate of pages loading in 1 to 3 seconds. Every 100 milliseconds you shave off lifts sales about 1 percent. Slow sites have roughly double the bounce rate of fast ones, and customers remember a slow site. They do not come back to check if it improved.
Google and AI search reward speed too
Google confirmed speed as a ranking signal. AI Overviews and other AI search results favor fast-loading pages, and mobile performance is the deciding factor since Google indexes your mobile site first. A slow site ranks below a fast one, all else equal.
The three numbers Google actually measures
These feed both search rankings and how AI systems judge a source. Understanding them tells you what to prioritize.
How fast your main content shows up. Under 2.5 seconds is good. If it takes 5, the visitor feels like they waited 5 seconds, even if the page was technically ready sooner. Fix: speed up the server response, load the visible part first, cut anything that blocks rendering, and serve content from somewhere closer to the visitor.
How fast the site responds to a click. Under 200 milliseconds is good. A page that looks fast but ignores a click feels broken, and customers hate it. Fix: delay anything not essential to the click, break up long-running scripts, and speed up the database queries behind interactive features.
Whether things jump around while loading. Under 0.1 is good. Customers hate clicking a button and having it slide out from under their finger before the tap lands. Fix: reserve space for images before they load, and avoid injecting content above what is already on screen.
The 10-point fix checklist
Not every fix is equal. These ten typically deliver the most for the least effort. Hand this list to whoever maintains your site.
- Compress your text. A one-line setting change. 60 to 80 percent smaller files, free, five minutes.
- Compress and modernize your images. Images are 60 to 80 percent of most pages' weight. 50 to 80 percent smaller, free, needs some tooling.
- Only load images the visitor can actually see. 40 to 60 percent faster above-the-fold loading, free, one line of code.
- Strip out unused code. 30 to 50 percent smaller files, free, handled by most modern build tools automatically.
- Delay anything that is not essential immediately. Chat widgets, ads, analytics. 40 to 70 percent faster initial load, free.
- Serve content from somewhere close to the visitor. 30 to 50 percent faster for anyone far from your server, $0 to $200 a month.
- Fix your server's response time. Often the real bottleneck. 20 to 60 percent faster, cost depends on the cause.
- Tell browsers to remember what they already downloaded. 80 to 95 percent faster for returning visitors, free.
- Reserve space for images before they load. Fixes the jumping-around problem, free.
- Remove tools you no longer use. 10 to 30 percent faster on a bloated site, free, and requires only an audit.
Sites shipping 14 marketing scripts on the homepage are the rule, not the exception. Each one arrived with a meeting where someone said "it is only one tag."
Real case study: Cuez, a product that got 10x faster
Cuez is a platform for broadcast and live-event production, scripts, rundowns, timing, and media all in one place. Response times were degrading the experience during live productions, where every second counts.
Before: 3-second responses. Users reported lag during live shows.
What was actually wrong: the database was asking the same question over and over instead of once, nothing was being remembered between requests, key lookups had no shortcuts, and some custom code was doing what the framework already handled better.
The fixes: eager loading cut database trips per request from 100 plus to 2 or 3. A caching layer stopped repeated questions from hitting the database at all. Adding shortcuts to the database dropped query time from 400ms to 50ms. Framework-native tools replaced homegrown code.
After: 300ms, a 10x improvement. Infrastructure cost dropped about 40 percent. The system now handles 10x more people at once on the same servers. The speed complaints stopped.
The full breakdown lives in the Cuez API optimization case study. For a related case, see Imohub, where 120,000 plus listings needed sub-0.5s responses.
What this costs and what it is worth
Speed work scales with how complex your site is and how much is wrong.
| Scope | Cost | Typical yearly return | Timeline |
|---|---|---|---|
| Quick wins (caching, compression, images) | $2K to $5K | $30K to $100K | 1 to 2 months |
| Standard fix (content network, database, code) | $5K to $15K | $100K to $300K | 3 to 6 months |
| Full rework (architecture, framework upgrade) | $25K to $100K | $300K to $1M plus | 6 to 12 months |
A typical calculation: 10,000 monthly visitors, 2 percent conversion, 200 customers a month, $100 average sale, $20,000 monthly revenue. Cutting load time from 4 seconds to 1.8 seconds typically lifts conversion around 15 percent, or 30 more customers a month, $3,000 more a month, $36,000 more a year. Cost: $8,000. Payback: about 2.7 months.
Free tools to check your own site
You do not need to hire anyone to measure this yourself.
Google PageSpeed Insights (pagespeed.web.dev): tests against the three numbers above, scores 0 to 100, with fixes ranked by impact. Test mobile and desktop separately. Mobile matters more.
GTmetrix (gtmetrix.com): shows exactly what is slow, with a visual timeline of the page loading.
WebPageTest (webpagetest.org): the deepest option, with a video of the page loading on real devices and real network conditions.
Chrome DevTools: open any page, press F12, go to Performance, record. Shows exactly where the time goes. What developers use when they want the full picture.
FAQ
What counts as a "good" load time?
Under 2.5 seconds for your main content to appear. Faster is always better, 1.5 to 2 seconds is excellent.
How much does fixing this actually cost?
$2,000 to $15,000 for standard fixes, the top 10 checklist above. A full rework runs $25,000 to $100,000, and is rarely the right first move.
Will fixing speed break anything?
No. Speed work tends to improve code quality, not damage it. The real risk is doing it yourself and breaking something by accident, which is why working with someone who has done this before is the safer route.
How often should I check my site's speed?
Monthly at minimum. Speed drifts as new code lands. A deeper quarterly review is a good complement.
Will this actually help my search rankings?
Yes. Google ranks fast sites higher. Changes usually take 2 to 4 weeks to show up in rankings.
Does mobile speed matter more than desktop?
Yes. Google indexes your mobile site first, and roughly 70 percent of traffic is mobile. Fix mobile first.
Does speed affect whether AI tools cite my site?
Yes. AI Overviews, ChatGPT search, and Perplexity favor fast, well-structured pages. Slow pages get crawled less and cited less. The same fixes that help Google rankings tend to help AI visibility too.
Reflecting on speed work after 250 plus projects
After 17 years and 250 plus projects, the pattern is consistent. Speed problems are rarely one heavy image or one slow database question. They are ten small decisions that each looked harmless on its own. A library added "for this one thing," an analytics script that turned into seven, a database question that worked fine on 100 rows and broke on 100,000.
The work I find most satisfying does not look like progress in a screenshot. Cleaner code. Fewer tools. One well-placed shortcut in the database. The site loads faster, the team ships faster, and the hosting bill goes down. None of that photographs well. All of it compounds.
If your site feels slow, the first move is not to rebuild it. The first move is to measure it. Almost every engagement I take starts with a free speed test and a 30-minute call about what it shows. The answer is usually shorter than the owner expected, and the cost is usually smaller than they feared.
For deeper architectural fixes, see the DevOps guide, which covers caching, database work, and content networks. For the underlying rendering decisions that affect speed before any of this, see SSR vs CSR performance.
What to do next
Every second of delay costs about 7 percent of sales. The checklist above covers most of what slows a site down, and it is relatively cheap to fix.
This week: test your site on PageSpeed Insights, free, 5 minutes. Write down the three scores.
This month: ship the top 3 quick wins: caching, image compression, and loading images only when needed.
This quarter: if the score is below 75, get a full audit and a prioritized plan.
If you want a professional to handle it, I have shipped speed work on dozens of sites including the Cuez case above. I prioritize by impact over effort and tell you what is worth doing now versus later. Get a quote in 60s on the contact page, or see the website service for a fixed-price project.