Maybe people are going at this from the wrong angles.
You need to handle a couple million hits a month. Let’s do the math. Instead of “hits” which is nebulous, let’s just say actual valid visitors, or “people”, 2 million a month. If you divide 2 million down to minutes it’s about 45 visitors a minute. Doesn’t sound so crazy when you put it that way. Break it down further and it’s actually less than 1 visitor per second, a statistical 0, lol.
In any case, handling traffic, ak “visitors” is only somewhat based on abilities of the CMS, but also on the power of the server. You can put the best CMS around on a shared hostgator server and it still won’t handle 45 people a minute.
So first make sure your hosts can handle the traffic. The CMS is a secondary concern. How much actual bandwidth do 2 million users a month use? Are you hosting text pages or are they watching locally-hosted videos or downloading PDFs? What is the average visitor’s length of time on the site? How many pages do they view before exiting?
In any case, a good server is the first decision, to handle the actual bandwidth and connections. Apache itself will have limits, as does MySQL or nginx or whatever it may be. So first research some actual numbers. The “scale” you need to handle may have more to do with horizontal scaling, separating servers, clustering, load balancing, than the CMS. Picking a CDN may be more important than picking a CMS in some cases!
Hostgator, for example, literally says their shared hosts can handle 7,000 - 8,000 visitors a day. You need about 70,000 a day. This is server talk first before anything else.
That said, also ignoring the CMS, the hosting software is the next decision. Apache, for example, will handle everything you throw at it until it runs out of some resource like RAM, disk space, network bandwidth, or CPU. However, for concurrent connections, you have to look at settings like maxclient and adjust as needed. You can increase numbers and Apache will eat up RAM and CPU until everything halts. RAM is the big limiter here, not Apache.
Ok so now going beyond that. You’ve got servers with enough bandwidth, CDNs, horizontal scale, load balancing, whatever is needed. Then you’ve got your hosting stack, for example Apache/MySQL. Now it’s time to look at CMS.
The ability to cache well is a huge deal. The ability to minify. The ability to put out pages with small memory footprints. The ability to reduce queries. These help when comparing CMSes because even if Apache handles all the connections, we still have to wait for the CMS code (i.e. PHP) to assemble everything, query data, parse scripts, compile templates, and actually serve the page. Wordpress typically ends up on the slow scale without a ton of optimization. It can be made faster, by those in the know. There is one particular Drupal user (just random googling here) about 6 years ago, complaining about number of queries. In fact a lot of Drupal complaining is performance-related. Anyway, he had about 10k visitors a day (you might have 70k remember), and they clocked 6.1 million queries in just 12 hours. What if you need 7 times that, plus the other 12 hours of the day, for 85.4 million queries a day? yeeeeesh!
You mention “enterprise”. This is a nebulous term that doesn’t have specific meaning. Does it mean high traffic? Tons of sites have that and aren’t enterprises. Does it mean using a lot of bandwidth? na. Typically what they mean is that they want to use software that comes with warranty, support, backed by profitable businesses. This pretty means every CMS maintained by some guy in the Netherlands on Github is out of the question. This is why WP and Drupal are mentioned, typically because of support options.
Enterprises want exceptional security controls and extremely tight hold of things like updates. No update should break their websites, if in fact the enterprise software even updates very often, which they usually don’t. WP can update about every 20 days, not to mention all the plugin updates, this makes enterprise IT soil themselves. Everything has to be tested all over again, is it compatible with all our plugins? Will it break our templates and customizations? etc etc.
So your choices for CMS are now limited to actual companies with a good track record, solid and secure software, support availability, and track record of success.
Now look at the fact that you have a team of developers and probably a lot of custom needs such as the multi-language stuff (this can limit choices!). I would stay away from any CMS that focuses on non-developers and tries to abstract everything into wizards and drag-n-drop automated modules. You need control over all code and all parts of the template. The “easier” the CMS gets, the more abstracted you are from core stuff and ability to customize.
So take a look at a popular CMS like ProcessWire. I’ve used it, it’s cool, gaining popularity. But it’s pretty much owned by this dude Bryan Cramer. And support is “go ask in the forums” and if something is broken “hey do some pull requests and fix it”. This just isn’t what a large organization wants.
Speed tests seems hard to come by, people used to test and compare CMSes a lot more some years ago. But anyway, I found one comparison from some years back which showed a pretty sharp decline in page speeds and delivery once they hit about 50 concurrent users. 50 doesn’t sound like much, but out of your 45 visitors/visits per minute, you can bet some good caching and optimizing will be in order no matter which CMS you choose.
I’ve only used PHP and MySQL based CMSes, so I’ve got nothing to offer in the .NET/ASP or Java space. There are some Ruby projects out there too.
I would say take a look at MODX. I’ll be writing on it soon. It is more of a framework than a CMS, but it does both. There is a lot of freedom for developers to code and do whatever they want, with paid support options.
You already know about all the other popular ones, so I’m just throwing it out there to look at.
Good luck with you!
