(Disclosure: Microsoft is a client of the author.)
One of the fascinating things about the technology market is how it goes in circles and not only forgets the past but misses the point. For instance, let’s take the recent mega trend: PCs as a Service, or as we’ll call it, “PaaS.” Most think this is brand new, but it is in fact an attempt to return to what we had with the mainframe. An appliance on our desk and like it was back then the hardware is, and should be, largely transparent and, I expect, as these services mature they hardware will change dramatically and alternatives like Windows on ARM will start to look more attractive.
What is also interesting is that just about the time the last large mainframe/terminal implementations were being yanked out by their roots, IT started demanding desktop computing as a service again. This was back in the 1990s, and now about two whopping decades later, major OEMs are putting programs together to make it happen and reinventing the wheel to make it happen.
What Is PaaS really?
What is being requested is desktop computing as a service, but because this request is being filtered by the PC vendors they are seeing this as PCs as a Service and creating a huge opportunity for firms like Amazon. You see much like Amazon did with cloud, they could skip the interim step and move directly to an optimal solution which might be a higher-end version of Fire TV or industrial Kindle coupled with a cloud service that provided desktop computing at a massively lower cost.
What the market is asking for is something that is billed like a utility that provides a localized computing experience that adjusts automatically with staffing changes and can be easily scaled up and down like terminals once were. PCs are in the mix largely because there isn’t yet a viable alternative being pitched, and unlike the terminals of old, these new ones will need to be predominantly mobile.
Hardware/services evolution: Looking at past models
However, it is certainly possible that very quickly the employee’s smartphone could become the default device passing through a Windows experience to something with a larger monitor and a keyboard which is why I think there is a play here for Windows on ARM.
You see once you deliver any computing service at scale cost reductions on hardware flow to the vendor’s bottom line and you need high standardization to keep breakage and service costs down. Existing and previous industries that matured this model include large scale telephony, mainframes (initially, before the hardware was sold instead of leased), and cable. I think Telephony provides the strongest model for profit because it introduced the concept of “moves, adds and changes.”
What the telephone market discovered was that they could sell solutions at or below cost and then bill out for moves, adds and changes at a premium and that is where their profit came from. Quickly the base costs for the implementation were built into corporate overhead, and if a department or employee wanted something more they only saw the cost delta and no longer saw the base line costs. This allowed the vendors to sell upgrades at high margin while the customers only saw the delta in cost between the baseline solution and the upgrade making the result look like a good deal while the vendor received a nice margin increase.
Granted this market also showcases a danger because the PBX vendors, and that is the part of telephony I’m talking about, mostly went under when customers stopped leasing their hardware, bought it, and then refused to upgrade or replace it. If the PC OEMs cash cow this segment once a significant number of customers are on these PaaS programs and the customers buy their aging hardware instead of replacing it, this industry could drop into obsolescence much as desktop telephony did.
Keeping it interesting
Now the lesson unlearned from both the Mainframe and Telephony market declines is that, as a vendor, you must keep this interesting. In fact, we are watching the cable companies, as people cut the related cords, learn the same thing. You can’t depend on lock-in and treat the customers poorly or fail to advance the technology. This is because, if you do, the market will move, and you’ll end up wondering where your customers and, eventually, your market went.
Microsoft is run by Satya Nadella who isn’t a desktop computing advocate: he is a cloud advocate. Microsoft is already working to turn Office into a service and it shouldn’t be long before they realize that Windows needs to go the same way. If they could, rather than giving their OS to the OEMs to image and deploy, provide a much more consistent experience by turning Windows into a service that Microsoft would manage, they would likely reduce dramatically both their support costs and the complaints surrounding their platform (along with much of the breakage). This would also allow them to more aggressively pass out updates (seen now as critical to Microsoft better assuring security) and were it not for a lack of a viable versions of Windows for the phone I think they would have already been well along in this effort.
This direction is further supported by Microsoft’s increasing reliance on web services as their primary revenue source and the foundation for their growth. This is fueling massive investments in ever bigger data centers and, I expect, that eventual move to a Microsoft Subscription desktop. Once they make that flip, you must wonder what Microsoft will do with Surface. Maybe nothing, but…
We are approaching a bigger change than we realize, and the PC OEMs seem poorly positioned to take full advantage of it. I think this is providing a huge potential advantage to firms like Amazon to step in and pull an iPhone like upset and that the only thing that may save them near term is that Amazon got burned badly with Smartphones and may be unwilling to take this risk. However, their Chinese counterparts aren’t that timid suggesting the next big subscription wave could come out of China and be accidentally backed by Microsoft’s anticipated pivot to Windows as a Service, or WaaS.
What amazes me is how badly the market wants to move back to the original mainframe hardware model and hard the industry is trying to see the similarities. One thing is sure, this transition will be very interesting.
This article is published as part of the IDG Contributor Network. Want to Join?