Windows 8, with its new, dramatic Metro interface, is a huge gamble for Microsoft. A huge gamble not only because they risk slowing update cycles and attrition of their desktop installation base but also because the Windows desktop is an underpinning of the Microsoft ecosystem – one that can easily unravel if Microsoft fails to maintain a strong foundation.
As a technologist I have been watching Windows 8 for some time having been using it, in some capacity, since the earliest public betas. I’ve long struggled to come to terms with how Microsoft envisioned Windows 8 fitting into their existing customer base but have been, more or less, hopeful that the final release would fix many of my concerns. When Windows 8 did finally release I was, sadly, left still wondering why it was so different from past Windows interfaces, what the ultimate intention was and how users were going to react to it.
It didn’t take long before I got a very thorough introduction to user reaction. As a technology consultancy we tend to move quickly on new technologies and trends. We may not deploy beta products into production but when new products release our update cycles are generally almost instantaneous. We need to run the latest and the greatest all of the time so that we are ready for problems before anyone else allowing us to stay ahead of our customers. So Windows 8 started getting prepped for rollout pretty much on the day that it was released to manufacturing. This is when management got their first chance to try it out before the actual deployments started – the IT department had been playing with it since early betas.
Management came back to IT to ask critical questions concerning efficiency, usability and training. Their reaction was that Windows 8’s interface was confusing and highly inefficient requiring a disruptive “jolt” of leaping to and from full screen menus that caused mental context shifting and loss of focus. Many tasks require power user levels of knowledge to be usable while the interface seemed to be designed around low end “consumer” use and not very appropriate for people with the level of knowledge necessary to make the system functional.
It wasn’t that Windows 8 was unusable but failed at delivering the value traditionally associated with Windows, the value that causes us to traditionally move from version to version more or less without thinking and that is that sticking with Windows on the desktop delivers a predictable user experience requiring little to no retraining and an overall efficient experience. Windows 8 requires extensive retraining, makes workers less efficient even after adapting to it and expects traditionally casual users to need to be power users to be effective. While sticking with Windows is the obvious choice for IT departments with deep investments in Windows knowledge and skills (and tools), the value proposition for end users does not have the same continuity that it has in the past.
We read many reviews and consistently the answer to whether Windows 8 would deliver value to other organizations seemed to be focused on it being “good enough” and that with extensive training and all end users learning to “deal with” the interface issues and to learn totally new skills like jumping back and forth between mouse and keyboard, memorizing shortcut keys, etc. that the system could be made to be functional. But never good, never ideal. All concerns around Windows 8 aren’t about showing why it is better, just making it acceptable. Hardly a position that we want to be in as an IT department. We want to deliver solutions and value. We want to make our businesses more efficient, not less. We want to avoid disruption, not create it.
We even went to far as to visit Microsoft at a trade show putting on a display of Windows 8. Even Microsoft’s own staff were unable to clarify the value proposition of Windows 8 or even in their demonstration environment get it to work “as intended”. It is clear that even Microsoft themselves are not confident in the product or sure how their customers are expected to react to it.
The decision was made quickly: management wanted a demonstration of a Linux desktop immediately. The first test was Linux Mint which ended up being the final choice as well. The non-IT users were really impressed with how easy Linux Mint was to use for people with a Windows background and nothing else. It required no training – users literally just sat down and started working, unlike on Windows 8 where users were confused and needed help even with the simplest tasks like opening an application or shutting down the computer. And there was essentially no pushback, people were universally excited about the opportunities that the new platform could provide, whereas people were actively concerned about how painful working with Windows 8 would be both up front and down the road.
That Windows 8 blundered so dramatically as to cause a competing product to get auditioned was not that surprising to me. These things happen. That the reaction of the non-IT staff was so dramatically in favor of a Linux distro was quite surprising, however. Staff with no Linux exposure didn’t just see Linux as a low cost alternative or the lesser of two evils but were downright excited to use it. Windows 8 caused Microsoft’s worst fears to come true – using Windows is no longer something that users can choose because it is familiar and comfortable. If they feel the need or desire to test alternatives Windows will no longer compete on a “devil we know” basis like it traditionally has in the past but will need to compete on a basis of usability comparisons as Linux Mint, in this case, actually felt far more familiar and comfortable than Windows 8.
What did truly surprise me, however, was the ripple effect that changing the operating system had on the computing infrastructure. Because Windows was being replaced this caused a series of questions to arise around other technology choices. The first, probably somewhat obviously, was what to do about Windows-based applications that had no Linux versions?
We are lucky that the shop ran very standard applications and most applications are modern, browser-based ones so the bulk of systems worked on Linux transparently. The only major application to require an alternative was Microsoft Office. Fortunately the fix was easy, LibreOffice had everything that we needed and is built into the operating system. Moving from MS Office to LibreOffice can be simple or intimidating depending on outside dependencies, complexity of use scenarios, heavy use of macros, etc. We were lucky that across the board the move was trivial, in our case.
Dropping Microsoft Office left us without an effective email client for our Exchange email system. So again, management asked, what compelling value is there for us in Exchange. Shoulder shrugs followed. Almost immediately a migration effort from a hosted Exchange service to Rackspace Email began. This resulted in one of the largest cost savings, overall, in this entire process.
Next to be questioned was SharePoint. Without desktop Active Directory integration, Microsoft Office integration and Exchange integration, was the overhead of running a heavy SharePoint installation of appreciable value to our organization? SharePoint put up the biggest fight as it truly is a nearly irreplaceable system with numerous aspects and features that cannot be trivially compared to other systems. In the end, however, without the slew of Microsoft integrated components SharePoint was deemed too costly and complex to warrant using on its own in our environment.
One by one, Microsoft products whose values were established through their tight integration with each other began to be eliminated in favor of lower cost, more flexible alternatives. As one by one they were removed the value that they had cumulatively created diminished making each one less and less valuable without the others.
Before the move to a Linux desktop we had been preparing to install Lync as a replacement both to our instant messaging platform as well as our telephony platform. Needless to say, that project was cancelled and our current systems, which integrate really well with Linux and were of much lower cost, were kept.
As we got to the end of eliminating Microsoft-based applications it became apparent that using Active Directory for centralized authentication was not cost effective. This last piece will take quite some time to phase out completely as creating a new, centralized authentication mechanism will take quite a bit of planning and implementation time, but the process has begun to move to a completely different platform.
Even applications that we thought were sacred and untouchable, where plans were in place to keep them running on dedicated Windows instances just for special purposes like accounting, ended up being less sacred than we had anticipated. New applications were found and systems were migrated.
Of course support infrastructure followed as well with System Center and Windows-focused backup systems no longer needed. And Windows-based file servers stopped making sense without Windows clients to support.
At the end of the day what was so shocking was that the littlest thing, a concern over the efficiency and usability of Windows 8’s new interface, triggered a series of discoveries that completely unraveled our Microsoft-centered ecosystem. No single product was unloved or disliked. We were a team of dedicated Windows 7 desktop users on a wholly Microsoft infrastructure and happy with that decision and happy to be continuing to move more and more over to the Microsoft “way”. But by simply questioning the assumption that we wanted or needed to be using a Windows desktop ended up bringing down an infrastructural house of cards.
From an end user perspective, the move to Linux was effortless. There has been quite a bit of retraining and rethinking from the support side, of course. There is a lot to learn there, but that is IT’s job – support the business and do what needs to be done to make them able to work most efficiently.
Does this bode of a dark future for Windows? Unlikely, but it does highlight that a significant misstep on the desktop platform could easily put Microsoft’s market position on a downward spiral. Microsoft depends on tight integration between their systems to create their value proposition. Losing the desktop component of that integration can quickly undermine the remaining pieces. To be sure, ours is a special case scenario – a small firm with extensive UNIX skills already existing in house, an ambitious and forward thinking management team and the agility to make broad changes combined with more than a decade of seeking platform independence in application choices, but just because we lie on the extreme edge does not mean that our story is not an important one. For some, Windows 8 might not only represent the tipping point in the Windows desktop value proposition but the tipping point in the Microsoft ecosystem itself.
2 thoughts on “The Ripple Effect of Windows 8”
Very interesting writeup. As an IT professional, testing W8 a while back left me scratching my head as the product was about to be pushed to a final release that they would magically pull the trigger and have a true desktop mode (complete with Start Menu) and mobile/table mode which makes sense in that space. Sadly MS has really gambled that their foothold in the markets would be able to weather the user reaction, and hoped that “deal with it” would be satisfactory enough.
As of right now I am just in the middle of a XP to W7 migration at my new business, so there is still time for MS to redeem itself. If it were not for the lack of critical application on the Mac platform I would have pulled the trigger on new Minis. But we are still to far out to phase out critical CRM/SCM/ERP solutions with only thick Windows applications. Two years down the road, who knows.
I read this and started wondering how I could start moving our agency in this direction. The major hurdle for us would be running some Windows apps on a linux machine. Then it occurred to me that with RemoteApp on a Windows 2008 or 2012 server, those apps could be piped to a linux client. And THEN, I found that someone had already worked this out…
It’s not a total departure from Windows, but it would vastly reduce the client OS footprint. And that’s a damn good start.