No Upgrades? You’ll Scare Your Users Away

Posted on by | 17 Replies

Today is Halloween so it’s a good time to addresses a scary topic.

After a baby-hiatus this summer, I’ve been back on the road again teaching a lot of Joomla and Drupal. I’m talking to a lot of end-users and finding that there’s something really scary In our open source platforms.

No Upgrades? Your Customers See an Exit Door

I tweeted a while ago that a lot of people are migrating to Joomla and Drupal from the Vignette / OpenText system. Why? Because OpenText are dropping support for an old version and moving to the new version requires a major migration. People are using that opportunity to shop around. If they’re going to have to rewrite their whole website, why not see if there’s a better alternative out there? Every time there’s a difficult upgrade, your customers seen an exit door.

The worst part of every class with Vignette users is explaining that neither Joomla or Drupal are better with upgrades. It’s not comforting to explain to new users that they got out of the frying pan and into the fire.

Times Have Changed

Here’s the thing: people’s expectations have changed. This isn’t 2005 when phones never updated and new Windows versions often required buying new computers.

In 2011, your phones, computers and tablets update automatically. In 2011, WordPress runs 15% of the web and its updates are like butter. In 2011, we all run on the cloud and we never even notice the updates. That’s today’s standard. Any system which doesn’t provide an easy update path isn’t living in the 2010′s.

Conclusion

I hope that Joomla devs provide a smooth upgrade from 2.5 to 3.0 and that Drupal devs provide a smooth path from 7 to 8.

I fear a scary experience for the users of either platform if they don’t.

  • OgyDog

    Steve, I agree with your post.

    If you have to do a major migration every time, at minimum, you should promise your client a website lifetime of 3 years. Migrating a site every 1.5 years or 6 months is not sustainable and does not provide any business ROI.

    For example, we just announced the end of support & maintenance for Joomla 1.0 sites in Apr 2012.

  • Dan Knauss

    You are correct. As soon as WP 2.9 and the iphone showed up, an old problem with Joomla and Drupal got a lot worse.

  • Steve Burge

    Hi Dan – not worse exactly, but certainly treading water for a while whilst WordPress advanced.

    I’m delighted that 1.6 > 1.7 > 2.5 is an automatic upgrade. I’m hoping strongly that will continue on to version 3.

  • Dan Knauss

    Definitely it’s a plus to have that in there. I hope it is adopted widely and effectively by quality extension developers, perhaps also with the effect that it weeds out the not-so-great ones.

    How are commercial extension developers that expire customer access to upgrades dealing with this new convenience? I would hope to see the emergence of a commercial support model that does not involve expiring access, at least not within a core LTS cycle, since this just increases security vulnerabilities and end user inconvenience, which ultimately reflects badly on Joomla. It seems there has been a tendency toward engineered inconvenience and obsolescence that is “cured” by the ol’ paywall. That drives good potential customers off not just because of cost or the existence of free alternatives but because the transaction is not based on trust and convenience. With SaaS and “the cloud” you’re paying a known monthly or annual fee, and all upgrades and improvements are made for you behind the scenes. As people get hooked on that model more and more, they will reject the idea of a self-hosted Joomla site that uses a bunch of extensions, each with its own upgrade paywall.

    I have done very little with J!1.6-1.7, certainly no production deployments, since it just hasn’t been worth it yet in the cases where I considered the option. My sense is that I’ll jump into 2.5+ only when it’s clear that total costs of operating and sustaining are competitive with the alternatives, which is often WordPress nowadays, especially if you favor the keep-it-simple principle.

  • Steve Burge

    Ah, now that’s a different question :)

    SAAS is a threat to change and diminish the role of most open source projects as they exist currently.

    Maybe the hassle-free aspects of the cloud are also part of what people expect nowadays. Time for most us to shift business models if so …

  • Ronnie

    maybe im way off but as of Joomla 1.6 and 1.7 upgrades seem to be one click.. just updated 1.7 to 1.7.3.

    Hopefully it stays the same once they go 2.0 on us

  • Steve Burge

    That’s what we’re strongly, strongly hoping for. Signs are looking good at the moment …

    A difficult upgrade from 2.5 to 3.0 will be terrible for our user base.

  • DanK

    So true. Maybe it’s just confirmation bias because it tracks my own views and decisions, but I have a strong feeling that the massive decline in brand strength, adoption and retention of users for Joomla is largely due to 1) the lack of value in costly upgrades/migrations between major releases that do not play well with extensions as well as 2) the maintenance costs and related insecurity issues that come from the lack of easy upgrades or simply centralized alerts for new releases of *extensions.*

    The history of long delays between significant releases would probably not be too bad in itself if not for these other things. Running a Joomla site has meant having no major core functional enhancements and reliance on a raft of extensions that, along with the core, require individual attention to keep them updated and playing well together. It feels like treading water at best.

  • Steve Burge

    I’ve mentioned elsewhere that I worry about Drupal for the same reason. I included them here for a reason. Their users are generally more technical but … the majority of site launches we do training for are still D6 because so much remains in beta – many big organizations won’t allow D7 even a year after launch. Now D8 is rolling with no prospect of an easy migration from 7.

    Whenever we teach either Joomla or Drupal, the worst moment of each class in when we explain previous (and current) upgrade policies.

  • Todd Chronister

    I work today in D6 in Q1 2012 after staring out in open source CMS with Mambo Q2 2004 and PHP since 2001. Thinking D7 in Q4 2012 but who knows – depends on the open-source modules I need for the project. My thought is that D6 is robust today – a veteran and predictable. I can sell that.

  • Ernest E Vogelsinger

    Steve, I agree with you on the baseline, however you cannot (yet???) compare either Joomla! nor Drupal with CMSes like OpenText or the ancient Vignette, for only a few reasons:

    1) For bigger websites (aka “portals”) a concise, manageable workflow ist missing badly.

    2) Again for the big ones, staging is essential. As of now there’s only one stage “CMS” that includes everything from development to delivery – effectively a no-go for the big ones.

    3) Finally, I couldn’t agree more with OgyDog as for commercially driven platforms an 18 months lifetime is ridiculous. They might accept an 18months release cycle, but need support for any given major release for at least 24 months after the development cycle end.

    I know that Joomla! and Drupal are OSP, so these points are simply yelling for some kind of commercial support offers, and issue 1) and 2) might ask for a commercial branch, or extensions.

    I’m not too big a fan of Drupal but got to like Joomla!, so if there’s interest from your side you might count on me.

    (Business background: responsible for the #11 portal in Austria – yes, we’re using OpenText :cry: )

  • Steve Burge

    Great comment, thanks Ernest.

    Have you come across any CMSs that offer really effective staging?

  • Ernest E Vogelsinger

    :sigh: uhm, not really, no. OpenText in fact does offer a workflow that actually works, and _in_theory_ could offer some sort of staging, but would not easily provide a clean separation of developemnt/integration/production, at least not to my knowledge. It would be nice if a transition between separated systems would be easier than a complete project export/import which is a chore from production to integration but absolutely impracticable the other way round ;-)

    What I do like with OpenText is the ability to expand the front end functionality, but that’s a feature that a separate Joomla! or Drupal delivery stage could easily provide.

  • Steve Burge

    I thought so. A few times now I’ve read the grumblings of people moving their sites back to flat file systems. One key complaint with the current database-driven CMSs is the poor staging support.

  • Ernest E Vogelsinger

    Actually I believe implementing staging, versioning and workflow on at least the content/media level could quite easily work with Joomla! as it offers a more or less clean DB design, plus a completely event driven system. Some small enhancements to the table definitions (basically adding a column for the staging level and another for the component/content version) would do the major trick.

    Adding a workflow, queueing, etc is effectively an issue of designing the right extensions, imho. I believe you could even create a multi-instance system for busy sites, where you have balanced CMS instances, a publisher instance that delivers content to staging hosts, and dedicated front-end instances to carry the user load.

    If done right a system like Joomla! v.10.01 would be able to compete in the big player business. The more as Zend is even more eager to plant its foot into the heavy load business.

  • pepperstreet

    Interesting article topic. Reminds me on the new Joomla 2.5 update feature. Smooth? It might be… if it would be usable.

    It requires “allow_url_fopen”, which is deactivated on serious hosting environments for security reasons. :cry: :o :sigh: :sad:

  • Amy Stephen

    This is my opinion on how the release cycle should work. I do not agree that there should be a migration from 2.5 to 3.0.

    The N.0 release is where all the change should be expected. This is the release for core improvements. Everyone should understand that the API will be broken. It should be understood that extensions will not work perfectly. This isn’t the time to migrate existing sites, it’s the time for developers to begin their work.

    While the middle releases might add new core features, these additions should be limited to finishing what didn’t quite make N.0. A commitment should be made to not break the API. These releases are for the extension developers to focus efforts on readying more and more extensions. During this period, new sites could be built with the release but it is still not time to migrate.

    The N.5 release cycle should be *purely* a bug fixing, stabilization and migration release. The focus should be on bringing along the user base and sanding the rough edges off. Absolutely no API changes. Absolutely no new features. Absolutely no data model changes. We should focus on the user base.

    I am hopeful that as Joomla gets used to the new release cycle, this is what will sort out. Each of these major focuses: core improvements, extensions, and users, matter and setting aside time to focus on those areas makes good sense.