How many visitors can shared hosting handle? A realistic calculation
TL;DR: There is no single number, because shared hosting is limited by resources, not by visitors. Typical limits are about one CPU core, 1 GB of memory, a cap on simultaneous PHP processes (often around 20) and disk I/O. A visit becomes several requests, and an uncached WordPress request keeps a PHP process busy for a fraction of a second. Using stated assumptions, a cached WordPress site can usually handle many tens of thousands of visits a month on shared hosting, and an uncached one comfortably handles around 100,000 until traffic spikes or heavy plugins appear. Stores, membership sites and sudden spikes hit limits much earlier. Work out your own peak load with the formulas below, turn on caching and watch your resource usage graphs to find out when it is time to move up to a VPS.
Why hosts do not give a visitor number
"Unlimited visitors" and "handles 100,000 visits a month" are marketing phrases. Two sites with the same number of visitors can have completely different loads:
- A static brochure site served from cache uses almost nothing.
- A WooCommerce store with logged-in customers and live carts generates a database query storm for each visit.
- A page with many plugins takes longer to build, so each visit holds resources for longer.
What a plan can really deliver is defined by resource limits. Our guide to hosting resource limits explains each one. This post turns those limits into visitor estimates you can adapt.
What actually limits a shared plan
Many shared hosts use a limiting system such as CloudLinux, which gives each account a slice of the server. Documented example defaults for a typical shared account are:
| Limit | Typical example | What it means |
|---|---|---|
| CPU (SPEED) | 100% of one core | The most processing time your account can use |
| Memory (PMEM) | 1 GB | Physical RAM across your processes |
| Entry processes (EP) | 20 | How many dynamic requests can run at the same moment |
| Processes (NPROC) | 100 | Total processes and threads |
| Disk I/O | 1 MB/s | Disk throughput |
| IOPS | 1,024 | Disk operations per second |
Your host's numbers will differ, so look them up in your plan details or panel. The estimates below use these example values.
From monthly visits to concurrent requests
You can estimate your own load in five steps.
- Visits per day = monthly visits ÷ 30.
- Peak-hour visits. Traffic is uneven. As an assumption, say 10% of a day's visits arrive in the busiest hour. Use a higher share if your audience is concentrated in a few hours.
- Dynamic requests per second = peak-hour visits × pages per visit ÷ 3,600. Count only requests that run PHP: pages, search, cart, login and AJAX calls. Static files like images are served separately and cost far less.
- Concurrent PHP requests = requests per second × average time to build a page.
- CPU use = requests per second × CPU seconds per request, as a percentage of one core.
Your page build time is easy to measure with your browser's developer tools (the "waiting" or TTFB figure) or with a tool such as Query Monitor.
A worked table
These figures are an illustrative model, not a measurement or a promise. Assumptions: 3 pages per visit, 10% of daily visits in the peak hour, 0.5 seconds to build an uncached page and about 0.3 CPU-seconds used per uncached request. The "10× spike" column shows what happens if a campaign or social post briefly multiplies traffic by ten.
| Monthly visits | Peak requests per second | Concurrent PHP requests | CPU of one core | CPU in a 10× spike |
|---|---|---|---|---|
| 10,000 | 0.03 | 0.01 | about 1% | about 8% |
| 50,000 | 0.14 | 0.07 | about 4% | about 42% |
| 100,000 | 0.28 | 0.14 | about 8% | about 83% |
| 500,000 | 1.4 | 0.7 | about 42% | over 400% |
| 1,000,000 | 2.8 | 1.4 | about 83% | over 800% |
What the table shows:
- Concurrency stays low, so the entry-process limit of about 20 is rarely the first thing you hit at modest traffic.
- CPU is the real limit. At about 100,000 visits a month with no caching, a spike can already push you past one core, and at 500,000 a normal peak hour uses nearly half a core.
- Slow pages make everything worse. If a page takes two seconds to build, concurrency and CPU rise accordingly.
What caching changes
Page caching stores the finished HTML of a page and serves it without running PHP or touching the database. A cached request can cost a small fraction of an uncached one, often a tenth or less, although the exact figure depends on your setup.
Using the same model with caching, where most anonymous visits are served from cache, the CPU numbers in the table fall sharply. A cached WordPress blog on shared hosting can often serve several times the traffic of an uncached one before limits matter. Our guide to what a CDN is and whether you need one explains how to take even more load off the server.
Caching does not help much with:
- logged-in users (members, customers, admins),
- carts, checkout and account pages,
- search results and filtered pages,
- highly personalised content.
Why spikes and bots matter more than averages
Your monthly average hides what causes trouble.
- Traffic spikes. A viral post, a sale email or a news mention can multiply load briefly. Shared hosting copes poorly with sudden bursts because of the hard per-account limits.
- Bots and scrapers. Search crawlers, AI crawlers, vulnerability scanners and scrapers can make thousands of requests, and each uncached request costs CPU. Check your raw access log. See blocking AI crawlers with robots.txt.
- Login attacks. Brute-force attempts on
wp-login.phpandxmlrpc.phpconsume resources. Our guide to securing the WordPress admin against brute-force logins covers rate limits and protection. - Background tasks. Scheduled jobs, backups and the admin dashboard's Heartbeat requests add constant load. See fixing high CPU from admin-ajax and Heartbeat.
Sites that hit limits early
- WooCommerce stores. Cart and checkout cannot be fully cached. See how much RAM WordPress and WooCommerce need.
- Membership and LMS sites. Most visitors are logged in.
- Sites with heavy page builders and many plugins. Each page takes longer to build.
- Large databases. Slow queries tie up resources.
- Forums and community sites. Lots of dynamic, personalised pages.
- Sites with frequent imports or API calls. For example syncing inventory.
For these, plan on a managed WooCommerce or VPS plan much sooner. See managed WooCommerce hosting in India.
Signs you have outgrown shared hosting
- You see "Resource Limit Is Reached" or 508 errors.
- Your site slows noticeably during busy periods.
- Your host warns you about CPU or I/O usage.
- Time to first byte regularly exceeds a second even with caching.
- Your resource usage graphs repeatedly touch their limits.
- You rely on tools such as heavy backups or imports that fail on shared hosting.
- Your host suggests upgrading.
Occasional brief touches of the limit are not a crisis. Repeated hits during normal traffic are.
How to test your own site
- Check the host's resource usage page (in cPanel, look for Resource Usage under Metrics). It shows CPU, memory, I/O and entry processes with limit hits, called "faults".
- Measure your page build time. Use your browser's network panel or a speed testing tool. Note the time to first byte for a cached and an uncached page.
- Run a gentle load test. A small test with a tool such as k6 or a hosted load tester, at a low number of virtual users, shows how response time grows with load. Ask your host first, because load tests may violate acceptable use policies on shared hosting, and never test a site you do not own.
- Watch real traffic. Compare your analytics peak hour with your resource graphs.
- Repeat after changes. Test again after enabling caching, updating plugins or changing themes.
What to move to next
- Better shared or managed WordPress plans raise limits and add caching at the server level.
- Cloud or VPS hosting gives you dedicated resources and control. See cloud hosting versus VPS and our guide to choosing a VPS plan in India.
- A CDN and object cache reduce load whichever plan you use.
Move when the cost of slowdowns, lost sales and troubleshooting exceeds the price difference.
FAQs
How many visitors can shared hosting handle?
It depends on caching, plugins and your plan's limits. A cached WordPress site can often handle many tens of thousands of visits a month, and an uncached one around 100,000 under steady traffic, but spikes, stores and heavy plugins reduce that sharply.
Is "unlimited visitors" real?
No. Every shared plan has resource limits for CPU, memory, processes and I/O. Visitor numbers are limited indirectly by those.
What is the first limit I will hit?
Usually CPU, or the cap on simultaneous PHP processes if requests are slow. Memory and I/O limits come next on heavy sites.
Does caching really help that much?
Yes for anonymous visitors on pages that can be cached. It does little for logged-in users, carts and checkout.
How do I know if I have outgrown shared hosting?
Repeated 508 or "Resource Limit Is Reached" errors, slowdowns at busy times, usage warnings from your host and resource graphs constantly hitting their limits.
Will a CDN fix it?
A CDN reduces load from static files and can cache pages at the edge, which helps a lot, but it does not fix slow dynamic pages.
How many concurrent visitors can I handle?
Concurrent requests depend on page build time and your process limits. If a page builds in half a second and you have 20 entry processes, you could in theory run about 40 dynamic requests a second, but CPU usually limits you first.
Should a small business site worry about this?
Usually not. Most small business sites with caching sit well within shared hosting limits. Worry about it when you grow, run a store, or expect big spikes.
Conclusion
Visitors do not break shared hosting. Expensive requests do. Turn your monthly traffic into peak requests per second, multiply by page build time and CPU cost, compare with your plan's limits and keep caching on. When your graphs and error logs show repeated limit hits, move up.
HostCloud's plans make resource limits visible and give you clear upgrade paths as your traffic grows. See the options at https://hostcloud.in.
Sources: