Managed WordPress Hosting Performance Benchmarks | Hacker News Reader

Managed WordPress Hosting Performance Benchmarks

reviewsignal.com

33 pointsandreyvit27 comments

Oh hey, it's my old company ASO on there! I haven't followed Doug and the team much since I sold it, but good to know it's still a contender and the support levels are still as high quality as I left it.

Did you have a chance to test on one of their business accounts? http://asmallorange.com/hosting/business/ It's like a standard shared account, but not oversold and therefore able to better scale.

This was tested on a VPS ($25 one) using their custom LEMP WordPress stack.

Good to see SiteGround keeps up, I've found their offering very compelling and signed up for them. Support does seem good so far, and I think they can compete on cost because they're mostly based out of Bulgaria from what I can tell. Their highest WordPress plan tops out at $15...so that's half of WP Engine's starting plan.

I am amazed noone had mentined Dreampress $20/month (includes two VPS files + mysql and uses varinsh cache) which they claim is equal to $500 wordpress.com VIP. I am using it, but I never had the privilege so far to hit more than 100 concurrent users.

They are on the list for the next round of testing.

Okay!

Please do Flywheel next time too

They are on the list too! Pantheon and Media Temple as well.

I'm curious how pantheon would have done in this test as it is the new kid on the block. https://www.getpantheon.com/wordpress

author here. If/when I get around to doing an update of these tests there are a lot of companies missing (flywheel, mediatemple, dreamhost, etc). Pantheon added it while I was working on this, so I didn't even get a chance to consider them for this go. I will definitely keep them in mind though.

Flywheel is amazing, you should really add them to the test.

http://getflywheel.com/

Next round, definitely

Could you please at least update this version of the article with a link to the demo page that you blitz'd? Or, at least one public link to a webpagetest.org result that shows the number of requests, grades and total amount of data downloaded per page requested?

This is a great writeup and a real gem among a sea of junk hosting reviews out there, but I am having a hard time couching these results against my current service because blitz.io, and webpagetest.org results obviously vary to a HUGE degree depending the site / page tested and how it has been optimized.

My regular testing includes webpagetest.org and blitz.io as well, so I just need a little more of a frame of reference to decide whether or not I want to use one of those coupon codes and make a switch. Thanks!

Let's see if this works.

EDIT: nope. That's was supposed to be a table of the webpagetest.org results :/ its 3am here and I need to get some sleep and be awake again in a few hours. I will have to get back to you on a nicer formatted version.

Thanks. That's actually more than enough to help me understand and compare. For completeness you may want to mention in your original review that you used a default wordpress install for the tests which is ~250-500kB downloaded and listed as "Bytes In" in a webpagetest.org result. Maybe you did and I missed it.

I started my tests with a default install too, but found that the actual sites I deal with are in the 1-2MB range so I imported "dummy content" I found somewhere to get more representative data and test different stacks of my own creation.

I decided to go with a $40/mo un-managed VPS 6 months ago partly because managed would have been hundreds of dollars more per month and partly because I wanted to learn how to do it myself. After approx 6mo of testing and optimizing various stacks with similar results to these managed hosts I have to say they represent a tremendous value. Configuration management, back ups, maintenance, security and disaster recovery of a high performance stack is a lot of work and I am not sure if I want to keep doing it.

That said, I could probably start my own managed hosting company that would compete well at this point!

I mentioned it was the default install but maybe that wasn't clear enough. I changed them all to 2014 theme too to give some parity. Real sites get heavier and heavier. When I load tests later they were on more realistic sites with some popular plugins enabled and dummy content. It's hard to make a 'representative' site though. Each install I've worked with has it's own set of plugins and feature-creep, it's hard to benchmark something like that and say it scales across everything.

It definitely is a lot of work managing your own sites and these managed providers certainly fit a niche. You wouldn't be the first to jump in and start offering it yourself. Pressable has a bunch of code shared on GitHub if you were serious about it too.

If you ever are interested in writing about your experience optimizing WP, I was thinking of doing an article like that as well. Maybe you could do a guest post? kevin at reviewsignal is my email if you're interested.

Interesting benchmarks. I am really impressed with the depth you have gone with these tests. It will be great to see newer companies included in that list too like cloudways.com and jumpstarter.io.

Pantheon, Cloudways and Jumpstarter are three great platforms with a potential to beat a lot of higher players in this industry.

I'd like to see this test run again... but with a lot more WP hosts than were originally included in the test.

It totally blows my mind that Wordpress powers (or claims to power) 19% of the web. I had never used it until this year (after using other CMSs like SquareSpace and Statamic) and wanted to blow my brains out. I really wonder how the CMS landscape will look in five years.

I hear people say stuff like this a lot. I've been using WordPress for years and find it really, really good -- not hard to get into, infinitely customisable, well supported, always improving, well-written plugins for every function that a reasonable number of people requires etc etc.

Is it just because it doesn't use the latest methods or languages or something? Is there really more to this than simple techno-snobbery?

FWIW I think the answer to your question is dead simple -- WordPress will still be the CMS leader in five years, running something rather greater than 20% of the web. You can't beat the entry price...

The major issue I've had with wp is the speed. I don't know if it's a php problem or a wp problem, but running wp on shit-tier commodity hardware results in a complete inability for your site to be {hn'd/slashdotted/dugg/whatever} due to the ~500ms page generation times (that only increase with any plugins at all). The caching plugins mostly resolve this, but still not well enough to be accessible during a wave of pageviews.

It seems like unless you're willing to pay for managed wp hosting (or fancy hardware), you can't expect more than 10 pageviews/s to be handled. This makes me sad.

You can up the performance pretty significantly with caching at various levels. My experience was the plugins don't do as much as something like putting nginx in front. http://reviewsignal.com/blog/2013/08/29/reverse-proxy-and-ca... what I put in front of it, handles reddit/hn/digg whatever just fine as far as I can tell.

Caching somewhere above WordPress seems to do wonders, things like Memcache, Varnish, etc.

Also, some of the companies I was comparing are charging <$10/month and outperforming some of the bigger brand names like WPEngine (SiteGround and GoDaddy for instance). I suspect if they gain traction with their products, they could push the price down on entry level managed wordpress hosting.

> well-written plugins for every function

This is what I think the main problem is. There are so many plugins and many (most?) are poorly written. Since most users don't know how to code beyond maybe some html/css, they just search for the functionality and install whatever matches. That can typically land you with a WP site running 10+ plugins, each with their own security / bug liabilities.

WP is fine by itself and there are lots of good plugins but like anything user managed, it gets blamed regardless of what's causing the problem.

People use Wordpress because by now there's a plug-in for everything; I contribute to a smallish blog about grant writing (http://blog.seliger.com) and do simple maintenance, and I use it because everyone else does and practically everything I've ever wanted to do has had a plug-in already written.

Joel Spolsky's "Bloatware and the 80/20 Myth" http://www.joelonsoftware.com/articles/fog0000000020.html is relevant here:

A lot of software developers are seduced by the old "80/20" rule. It seems to make a lot of sense: 80% of the people use 20% of the features. So you convince yourself that you only need to implement 20% of the features, and you can still sell 80% as many copies.

Unfortunately, it's never the same 20%. Everybody uses a different set of features. In the last 10 years I have probably heard of dozens of companies who, determined not to learn from each other, tried to release "lite" word processors that only implement 20% of the features. This story is as old as the PC. Most of the time, what happens is that they give their program to a journalist to review, and the journalist reviews it by writing their review using the new word processor, and then the journalist tries to find the "word count" feature which they need because most journalists have precise word count requirements, and it's not there, because it's in the "80% that nobody uses," and the journalist ends up writing a story that attempts to claim simultaneously that lite programs are good, bloat is bad, and I can't use this damn thing 'cause it won't count my words. If I had a dollar for every time this has happened I would be very happy.

Replace "word processors" with "blogging platforms" and the same basic thing is true today of Wordpress.

wordpress has a lot of features (and plugins) if you want something like a multi-user blog, the code base is what it is,but noone can deny that.Furthermore for non technical user it's like one of the easiest thing to deploy and supports a wide range of cheap hosting providers.

Furthermore neither Statamic nor SquareSpace are opensource => vendor lockin.

It amazes me that so many people find WordPress inadequate and/or intimidating even though it requires so little of the admin in order to setup while being incredibly powerful.

I've used Drupal and Joomla and I can't fathom why anyone would prefer those platforms over WordPress.

Thenagain, I've been using WP since version 1.0...so maybe I have been trained to accept a way of operating that isn't necessarily ideal. But I've never heard an argument for why WordPress sucks (asides from the source code).

I think you see a lot of negative reactions about WordPress on places like HN because hits a bitter-spot (anti sweet spot) for most web application engineers:

* You don't start interacting with it via code or text files -- you administrate it through a series of web forms. * Customization is done through plug-ins rather than composing familiar framework abstractions * If when you do decide to look under the hood, it's not organized like any familiar back end from the last 10 years and will take you longer than usual to figure out * Oh, yeah, PHP

It's pretty hard to argue with its success for people who aren't webapp engineers, though. And even from the engineering standpoint, after I had spent 2-3 years working on projects that used it as a basis for a platform, I came to a grudging (if tepid) respect for it.

Reply on news.ycombinator.com