Is Sitecore Worth the Cost? A Perspective

[ergophobe]

Like most of the web, we tend toward discussions of Wordpress, Drupal and maybe a few others (Dot Net Nuke comes to mind).

What Enterprise Costs
We don't see much discussion of the big enterprise systems. For the last 8 months, I've had a chance to work with Sitecore a lot as a user, not a developer (more on that in a second). Sitecore pricing is a complete enigma. They don't post pricing on their site and I've seen developers say they have trouble even guessing at a ballpark number until Sitecore comes back with a cost based on the project. But my understanding is that you're looking at $150,000 or more for a license.

So that puts it out of the range of hobbyists and small businesses. In my opinion, that should put it out of the range of medium businesses too.

Given the cost of the license, the developer pool is much smaller than, say, Drupal, which is the open source platform that tries hardest to go toe to toe with Sitecore. This means that not only is licensing expensive, but developer costs are expensive too.

Sitecore Selling Point

In theory, the advantages are that it is a complete "customer experience" platform with built-in personalization available and A/B testing and more. And this gets to why this is not a small business tool - not only is it insanely expensive to license and develop on, but if you are going to actually realize its full potential, you have to throw a lot of money at it. In my experience, to really make use of the personalization, you would need at least a few full-time people managing your website.

So it's not only going to set you back $200,000 to license and build on the low end, if you don't have $150,000 a year in labor to throw at it, you'll not use it to its full potential. For example, I heard of one site that set up the personalization features, but then didn't have the personnel to run them. So instead of three (or was it five?) personalized versions of each page, they had one version that they now pasted into the four other now identical variants (or something like that - the story is third hand and all I know is it was eating resources that could have gone to better things).

My Experience: A Cumbersome System and Often Buggy

My experience trying to maintain a Sitecore site is that it is cumbersome, everything takes time and there are few if any bulk operations available.

For example, if you want to unpublish six pages, it's a multi-step process to go into each page and unpublish it. If I compare this to Drupal or Wordpress where I can check six checkboxes and hit an unpublish button, it's rather annoying. And it's like that over and over for all manner of tasks. Bulk image uploading, if available at all (depends on context) is buggy and a hassle (it asks you for the alt text, but then ignores it and flags a warning on all the images you uploaded. I've encountered a lot of other bugs.

A few other pleasures

- worst WYSIWYG editor I've ever seen. Terrible tag soup. Come on guys, just steal from Wordpress... except of course you can't because then your whole CMS would have to go GPL.

- no cron hooks by default. It's a Windows system, so I'm using the wrong terminology, but what I mean is this. I can schedule publish and unpublish operations, but they don't do anything on schedule. They wait until the next time I *manually* do a full site publish (which you should normal avoid because it's resource-intensive) and, if during the *manual* site publish, a givne page has a date flag that triggers, it too will publish or unpublish. I'm sure there are add-ins to fix this, but yes, the default system is this bad.

- Resource intensive. Just to get the admin end to work, we had to upgrade the dedicated server to 16GB of RAM and I think we eventually doubled it again and it's still pretty slow. This server doesn't even serve pages. It's just the editing/content management server.

- poor URL abstraction - the URLs match the content tree and the content tree matches the URLs. End of story. That's not that big of a deal, but I feel like it's emblmatic of the system in general, but that may be the system as given to me, not the system a developer would see.

- Takes over your browser. Tons of client-side scripting so, for example, you can't cut and paste text from a Rich Text area without opening the editor. Since that requires locking the page, that generates a new version in our system (that's a choice, but the versioning is awful).

- search and content filtering are also sub-standard

- Experience Editor is cumbersome. You can build complex page layouts using the Experience Editor. I avoid it if possible because it's so slow and cumbersome to use. I held off a long time in that judgment because I thought maybe I was just used to Drupal Panels and was prejudiced. At this point, I have spent far more time with Experience Editor than with Panels and there hasn't been one time I've thought "Wow, I wish Panels worked like this."

- Expensive to develop on? Here I'm at a disadvantage, because though I've done a lot of Drupal dev and have contributed patches back to the project, I only know Sitecore as a content manager. But I asked the developer to do something on our site that would be simple in Drupal. In fact, I did it for a Drupal client of mine in an afternoon using Views. The quote from our developer: $15,000, budgeting 110 hours. That knocked my socks off. This was a small tweak having to do with sorting events and how to sort and display events after the event date. Like I say, this was a few hours in Drupal.

I could go on, but at this point I'm more interested in hearing from someone who is a Sitecore fan. I know they have a great marketing machine and they are also very good at spreading FUD regarding open source options, but it must also have its merits that justify the price tag.

What is it that Sitecore does well? Who should consider using it?

I'm not sure we have a membership who can answer that, but if you have used it and, especially, if you've built sites with it, I'd like to know when it's the best option.

[blend27]

>>> $15,000, budgeting 110 hours

That is actually not bad: 15000 / 110= 136 / hour, considering a fact that its is built in C# and has a huge codebase: [github.com...] probably includes a PMO, a QA tester & Developer.

This is where real full time developers.programmers can make a OK living.

And then again, It is for a different business model(Sitecore), different market, probably not for Dorothy who sells socks on EBay/Etsy and wants to have a "website" with the Blog attached to it.

>>>So it's not only going to set you back $200,000 to license and build on the low end, if you don't have $150,000 a year in labor to throw at it.

And that is coming off taxable income anyway. consider $80,000,000 for a year worth of project(meet/plan/code/test) with 30 external personnel for a New & Shiny "product" after a newly hired CIO used it in his previous Job at another company that had cash to burn.

As far as the slow code: Perhaps you got a bad install.

[ergophobe]

>>That is actually not bad

I completely agree with your comments about the *rate*. The rate was actually $140/hr and I was rounding. I only included the total amount because I know the rate and knew the total and was calculating the number of hours budgeted.

It's not the hourly rate that I found high, it is the 110 hours. As I say, I was asking for something I had done for a client on Drupal in less than 10 hours. If they had budgeted 20 hours and rounded up to $3000, I would have found that reasonable. Also, this is a change order, so they are already up to their eyeballs in the system and don't have a ramp-up. If it were a standalone contract with a new contractor, I could see budgeting 10 or even 20 hours just to get up and running with system, get it running on their own dev platform, look under the hood at whatever code the previous contractor had created, and then and only then giving an estimate for the actual work.

Here's the rub - we wanted this on multiple websites. They bid 15,000 for the first one, and 12,000 for each subsequent one. Again, I cannot imagine building or modifying a Drupal module for $15,000 and then needing to budget 80% as much time to roll that module out a second and third time.

And again, this is a first experience with Sitecore and because of the licensing, I can't just download it and learn how hard/easy it is to work with. That's part of why I wrote this - I'm curious if this is normal for Sitecore or a particularly bad experience.

>>consider $80,000,000 for a year worth of project

But it's a question of value. There are all kinds of ways you can waste money and this is not the most egregious, you're certainly right about that. One of the things that I see is that people "bucket" money, so if marketing has a million dollars to spend and maintenance has a million to spend, it often doesn't seem to occur to marketing that if you build that site for $80,000 instead of $200,000, then two guys in maintenance don't get laid off (or everyone gets another $1000/year or, and this is the reason for the disconnect, some guy at the top of the pyramid pockets an extra $120,000 and there's no benefit to anyone but him).

>> bad install

That is more than possible. It is a first experience with Sitecore. I myself have built really slow systems only to profile them and realize one seemingly minor bit was slowing down the whole works (I remember cleverly pulling image dimensions in real time using PHP and then profiling the slow page to realize that was taking 99% of the execution time).

Either way, I'm surprised at how powerful the server needs to be to run this install. But as you point out, I don't have the experience to say whether this is norma or an edge case.

[graeme_p]

How large is the site, and what does it do?

Come to that, what does sitecore do? Their website has lots of buzzwords but no clear explanation. It looks like a heavyweight CMS with built in analytics.

[ergophobe]

It will be a couple dozen sites when fully built out, many of which will share a server and some sharing of codebase. The sites themselves are small (100-300 pages) and are mostly pretty simple sites in terms of functionality. Mostly they simply serve up basic (albeit sometimes complex) pages. But there's no e-commerce component, no customer login, nothing like that.

Sitecore's big claim to fame is personalization. It's not a CMS, it's a Customer Experience Platform.

What does that mean? Three things, I guess
1. Yes, there is some built-in analytics, but at least as our site is built, they are pretty worthless compare to Google Analytics.
2. Built-in A/B testing. You can create a "test" that compares the new version of a page against the old version.
3. Personalization. This is their big thing. You can segment people according to various criteria (first visit, have visited page X, came in via an email link) and show them different content.

[graeme_p]

A/B testing is not all that difficult, so to justify the cost AND tying yourself to a proprietary platform (which always scares me) 3.needs to work really well and be something you actually need and use. Is three really better than getting someone to write custom personalisation using opensource libraries? $200k buys a lot of coding.

[lucy24]

$150,000 or more for a license

Lifetime or annual? Right away you can see that it's only cost-effective if you would otherwise have spent significantly more than $150,000 on hand-coding. And honestly for that kind of money I can't imagine not doing it in-house and keeping control. Doesn't the law of diminishing returns kick in at some point?

[ergophobe]

Pretty sure that's lifetime.

>> if you would otherwise have spent significantly more than $150,000 on hand-coding

You probably would in order to get all the Sitecore functionality, but, I'm not sure how many people use all the Sitecore functionality (we certainly don't). The personalization features are similar to Acquia Lift, which also doesn't have a straighforward price (at least not on their website).