I came across this report through from here. It states that 58% of DNN sites are using DNN 4. I find this surprising as the number is very high. It's not 5 or 6. It's 3 major versions old. So the reasons could be due to major customizations or modules and skins which haven't been updated to newer versions. 

I am not sure if this means that these sites are not interested in upgrading due to high costs or because these sites just work fine or the probability that some of these sites have their sights on moving to another CMS like WordPress, Drupal and Joomla. I do know for a fact that WP is easy to upgrade. I have used WP and there's a notice in the admin area whenever there's a new version of WP. All one has to do is click the upgrade button and WP does all the work. Assuming no customizations have been done which might get broken in upgrades, the upgrade process is seamless. The same process happens for WP plugins. It's even easy to upgrade all the plugins in one shot. 

These stats tell me there's a big market for upgrade services. I am also surprised DNN sites are very slow in upgrading. Once they are several versions behind, it gets more complicated & costly to upgrade. There are upgrade services who won't touch v4 & v5 anymore (well.. maybe if they are paid enough).

The one concern I might have is that if I decide to develop a 7.x only commercial module, the 7.x market might not be big enough, seeing most sites are on v6-. I am also interested if dnnsoftware has any concerns about these stats and/or if they are doing anything about this. Stagnating DNN sites are not a good prospect for a growing company and not good for the DNN marketplace.

I personally do not like upgrades. I do this with Windows and I would do the same for DNN. If you have a v4 or v5 site, I suggest you migrate the skin or start with a brand new skin and give your new site a fresh look and start with a fresh DNN installation and work things in. It's a good time to clean up files and remove the baggage. I would say the  performance improvements and the modules which work on newer versions only are worth the upgrade. Plus over time it will cost you less to upgrade to the latest version if you're not very behind in the upgrade cycle.

FYI the 3^rd party verification systems are terrible and not trustworthy – they write simple scripts to recognise fixed tags within web pages and use that to identify the product. Certain CMS’s such as joomla/wordpress have (had?) explicit meta generator tags which makes them trivial to identify (in fact wordpress even used to emit the version, which of course was very handy for hackers). DNN takes a different approach as we don’t emit identifying material so as not to compromise the security of our users. This does have the disadvantage that they can’t easily identify DNN – most of them in fact rely on the word “dotnetnuke” along with a version number being in the page title (which is what happened in dnn 1.x and 2.x). If you check some of them such as http://w3techs.com/technologies/details/cm-dotnetnuke/all/all they aren’t even aware of dnn 7 – and if you get them to scan dnnsoftware.com they don’t detect it as a DNN site. In short the results they show are wildly inaccurate and should not be used as the basis of any decision making.

DNN 7.3 does reduce view state, which is great, and every user will benefit from optimization of client scripts, resulting in faster page construction after it's been downloaded. Other changes may improve or affect performance, depending on your concrete solution.

There haven't been many changes to the database, which is fine. But if you are running a site with a large number of users, folders or content, you might want to run my turbo scripts with improved procedures (which I did submit to DNN Corp, but my JIRA tickets have been withdrawn for other reasons).