ToolBox -

I now see the difference between paying $5 or $10/mo for hosting vs paying more for WordPress specialists and fast, reliable support. Well worth my investment and I’m happy I switched.

– Dr. Joe Kravitz
www.dentalimplants.us

Sometimes “more expensive” means “gouging.” Sometimes you get what you pay for.

With WP Engine, it’s the latter. Because:

To build the features and benefits that WP Engine provides would cost you six times more than just hosting with us.

Now we’ll prove it.

Let’s say you’re running your own virtual server at $40/mo and see what it would take to improve that hosting situation to match the features and benefits of running on WP Engine.

The Cost of No SPoF

Your virtual server is a “single point of failure,” meaning a component which, if it fails, will completely bring down your blog. If the virtual software fails (common) or the underlying hardware fails (less common), your blog is down.

Actually it’s worse than that — if MySQL gets overloaded or Apache freaks out or a bad operating system patch is pushed or a DoS attack hits or any number of other things happen, your blog is down. If you’ve ever run a blog with a reasonable amount of traffic, you know this happens more often than is comfortable!

So let’s suppose you decide “no more single point of failure!” Now you’ll have redundancy and backup. What does that look like?

  • You need two servers, so any of the various failures described above could happen to one server, but you still have the other one running. Assuming you were paying $40/mo for a virtual server somewhere, now you’re paying $80/mo.
  • You need a separate database server because if the database is running on one of the web servers and that server goes down, the database is down and the other server can’t operate! Actually you need two servers, otherwise that’s another single point of failure. So these are another $80/mo. Or you could reuse the two web servers but expand their sizes so they can handle all this computation and traffic, so either way it’s another $80/mo, bringing the total to $160/mo.
  • You need a load balancer to distribute traffic between those servers, and automatically skip a server if it’s down. You might think this could be another $20/mo virtual slice, but then that slice becomes a new single point of failure! Instead, you need to use a load balancer service such as Elastic Load Balancing from Amazon or a managed system like the $99/mo active-active load balancing from ServerBeach. Let’s average out your options and call it another $30/mo, bringing the total to $190/mo.
  • You also need to keep the filesystems of the web servers synchronized, so that when you e.g. upgrade a plugin or upload an image it’s available on both boxes. Some people use rsync to do this, but they’ll also tell you this causes problems like broken images or even broken code when (for example) a plugin is upgraded on one server but not the other but they’re sharing the same database. The safer way is to use a shared filesystem like NFS, but this again requires another separate server and introduces another single point of failure. Call it another $40/mo, bringing the total to $230/mo.

Conclusion: The burden of ridding yourself of the downtime due to having single points of failure is 5x the cost of having one, vulnerable box.

In other words, more expensive than you thought! And we didn’t count the time-cost of implementation.

At WP Engine, we also have no single points of failure, and in fact we have more failover than is even described above! The difference is we amortize these extra costs over all our customers.

The Cost of Scaling for Traffic

That $40/mo slice might work fine for normal traffic loads, but get in the New York Times or TechCrunch or on The Today Show and it’s guaranteed to go down — at the worst possible time.

To plan for traffic spikes, you have to always have enough resources for the spike, even though normally most of those resources go unused. How much is that?

That’s harder to say. Just doubling the size of your slice (and doubling your costs) will not double capacity as any seasoned IT person will confirm.

Here’s an example: We have a customer who doubled the size of their traffic spikes, but had to go from a cluster of 3 servers to a cluster of 9, and still didn’t handle the spikes. They moved to a better-built cluster of just 4 servers with us and have never had a problem again.

What about the cloud? We have another customer who tried to increase their cloud-based virtual server count when they got more traffic. After going from 5 servers to 33 servers and still unable to keep up with traffic, they moved to a 4-server cluster with us and all the problems went away.

It’s not just “more servers,” it’s the architecture of that cluster. You have to spend 2x or 4x more for capacity alone, and even then you might need to invest in the server architecture itself.

The Cost of Good Security

“I won’t get hacked because I’m not a target. I’m not doing anything offensive.” That’s what most people think, and they’re wrong.

Almost all attacks are not targeted at you personally! Rather, they’re trying to get control of a machine so they can run software to do what they really want to do: send spam email or hide their identity as they attack the real target. They don’t know you, they’re just scanning IP addresses and domain names looking for vulnerabilities.

Everyone is constantly attacked. We block 10,500 attacks every day.

To achieve our level of security, here’s what you have to do:

  • Get a top-tier security firm to audit your machines and lock them down. ($2,000-$10,000 depending on complexity.) Get periodic reviews to make sure new code or software patches hasn’t opened new holes, or that new exploits haven’t been found.
  • Use a firewall appliance with DoS (Denial of Service attack) prevention. (Ours costs $200/mo per cluster.)
  • Continuously scan for known malware using a service like Sucuri ($65/year).
  • Don’t use shared hosting (every single shared hosting provider has been successfully hacked at least once).

The final cost will depend on what degree you’re willing to go though to not get on Google’s Malware warning list and not get your site taken down by hackers. Whatever it is, it’s not $0.

The Cost of Good Support

We have the highest number of WordPress experts on staff in ratio to customers than any other hosting company on Earth. That means you don’t have to post questions to forums or spend $75/hour for basic advice and assistance.

Whether it’s your own time looking for answers online or your own money, it costs you hundreds of dollars a month to not have trusted experts on the other end of an email.

Conclusion

If you don’t care about site speed, or downtime, or traffic spikes, or getting hacked, or getting WordPress advice, then you’re absolutely right to stay with a shared hosting or VPS solution because it’s cheaper.

But if you do care, there’s no way to get the depth and quality of service we provide for less money.

Although we may seem more expensive on the surface, it’s the least expensive way to get the best WordPress hosting on Earth.