Summary: The little secret that IBM just might want to keep secret. How a simple consultant can sell OS/2 as a service rather than a product and bypass the whole "OS/2 or Windows?" question entirely.

I need to make clear that this point off view is not appropriate for everyone. It is most appropriate for consultants doing fee services and would be appropriate for commercially oriented ISVs, but it may be of interest to others. With my background in OS/2 brand management for IBM, and as a general advocate for OS/2, I've toyed with the "if I must play the hand I'm dealt; if I must work with circumstances as they exist, can I be successful with OS/2?" scenario. I think the answer can be "Yes". Let me make this the first salvo in what should be a lively discussion.

First, I need a business solution because you need to be solving a business problem. In our environment, we call those applications. For my example I will use Aviar's "Oz" application, but many OS/2, DOS, and Windows apps would work as well. If you are not familiar with Oz (the abbreviation for ounce), go to their web site. Oz is the OS/2 follow-on version to the DOS Ounce of Prevention Software (OOPS). These two applications are "computerized maintenance management systems" or CMMS. Applications which allow an organization to list tasks which must be performed, schedule the tasks, track costs, labor, inventories of parts, trigger purchasing requests, generate work orders, check compliance with regard to inspections, document required maintenance checks, record actions taken, create schedules. In short, a very handy application for an organization which has to track and document the execution of many tasks, including maintaining their facilities.

John Urbaniak, president of Aviar, can tell you that his sales suffer because this application runs on OS/2. Each sale of Oz involves OS/2 and UDB. So, how do we get an organization to use and OS/2 application if they don't want to use OS/2? My answer is: don't sell the product. Don't force them to say "Yes" to a "Do you want to buy OS/2?" decision. Position the application as a turnkey service; all software, hardware, and support becomes part of the contracted service.

When you talk about an application, the first thing users are likely to ask is "What does it run on?" In terms of both hardware and software, they are thinking of cost and skills to support it, how it fits in to their infrastructure. Being known as an OS/2, WindowsNT, or Java application may actually become more of a factor than the feature set or price of the application itself. How can an organization with no OS/2 skills consider buying an OS/2 application? But when you discuss a service with a user, the focus is on what the service does, the feature set and how that relates to the cost of the service.

This avoids lighting several fuses. First and foremost, you side step the emotional OS/2 decision. Who has the time to persuade everyone who thinks they "need" Windows that they are wrong? And the entire sale becomes an operational expense, not a capital expenditure. In many organizations, retaining a service is a much easier process to complete than to actually acquire - to purchase - an asset.

Now, I'm going to bring in my "fee services consultant", whom I will call Fred. Fred makes a living installing TCP/IP networks for people, doing general technical support, taking on some contracts for computer and network activities. It's a good life. But Fred's excellent OS/2 skills are not providing the return on the investment he made developing those skills. This is not always true, but more and more folks are asking Fred to do things with other products and platforms. Fred spent time and money becoming an OS/2 Warp Server guy. He's good at it. He likes it. Fred can do great things with it. But it just isn't happening very much anymore.

Fred knows he can take OS/2, client and server, and build a bulletproof network that's simply a never-go-down, seven-days-a-week, 24-hours-a-day environment. He's got options to meet requirements. He can do mirroring, backup server failover (VINCA), there's lots he can do in the design to make this thing an industrial strength, bet your business, mission critical network. And now Fred is beginning to see a way to make this pay off for him.

The PC client is simply an application platform, a line of business applications, communications and messaging tool. With OS/2, Fred can build a superior network to deliver the application to the user. Now Aviar has provided him with a superior application. Fred's off to visit prospects in his territory, blue collar outfits with a large maintenance or "physical plant" department. He is going to present Oz to them as fully supported service from Fred's IT Services company. All the hardware, software, instruction, support, and so forth are part of Fred's complete CMMS service.

What is the user "buying"? Well, first there is access to the excellent function of the CMMS package. But past that the user is buying the ability to retain a service; to "outsource" if you will, all the computer and IT related issues. No user operations or user administration and support. No network management. No software deployment or concerns about obsolescence, upgrades, or product life. I can rattle off all the tasks associated with "total cost of ownership", but the key message is that the user is buying peace of mind. Yes, it's function, reliability, availability, simplicity, but also peace of mind because Fred going to take care of it all. And Fred can do this for less money than the user organization can because this is Fred's business. He is a specialist.

This reduces OS/2 to an embedded operating system running in an appliance. That appliance just happens to be the CMMS workstation. You know it's a PC. I know it's a PC. But I'm talking about marketing positioning. OS/2 is no longer an issue to the user. Fred sets this up for them, customizes the application and the workstation, then trains them. All fee services for Fred.

As to the CMMS application itself, Fred can charge a monthly fee per workstation, thus setting up a recurring revenue stream for himself. There could be a base service fee per month, augmented by the number of "seats". He can break it down by access to the application, support, and other features. It's an open book without any rules. In fact, Fred can develop multiple pricing plans as part of one proposal to a prospect. Let them spend energies figuring out which is the best deal for them. An old marketing ploy!

Fred will also be slanting this so he gets other opportunities. For example, with Managed Clients from Serenity Systems, the WiseManager, DominoGo, APACHE, Hethmon Brothers Mail Server, and the clients have PMMail, RelishNet, FaxWorks, and Netscape Communicator. Fred can create a functional intranet with complete e-mail capabilities and internal web sites. But think of all the buzz words he can deliver to the user as additional services to the CMMS function.

Fred sets up the network, the CMMS workstations, maybe he gets a couple of other clients out of it because of the intranet angle. He has a monthly service agreement with a minimum 6 month commitment, cancellation penalties, and lower rates for extended agreements. His costs are pretty much covered and he got some fee service dollars for setup, training and customization. Just how well will he do?

In the beginning, his beeper may go off constantly as users get familiar with the system. But, after awhile, the "beeps" will become few and far between. And that's the sweet part. Once the shakedown period is over we all know this will run like a rock. He's services are not required on a day-to-day basis. And he can dial in to the server to do user and application management, even problem resolution using a product like NetOp. All he needs is his beeper, cell phone, and lap top computer. He is taking advantage of what Warp Server does best. Run all day and all night, without interruption.

Those pesky details

For the sake of discussion, I oversimplified my example. But not very much. It's based in reality although I admit it hasn't been done, yet. Oz exists and can run on OS/2, including diskless RIPL, and WorkSpace on Demand. Serenity Systems, working with its development partner, TouchVoice, can provide all the hardware and software necessary, including IBM Netfinity Servers with OS/2 Warp Server SMP pre-loaded. Serenity Systems also provides diskless RIPL client PCs. With no moving parts, these machines are extremely reliable.

And spares or replacement client systems can be kept around. With WiseManager, featured in the current issue of VOICE, workstations can be set up in minutes. This may be one of the few reasons Fred would actually have to go to his customer's site. But try this: Fred leaves some spare workstations on the site. They aren't very valuable except as CMMS workstations because they don't have hard drives, diskette or CD ROM devices. On the box, Fred writes the "network id" of the built in ethernet card.

Now a client system fails and Fred gets beeped. Fred says, "Go to the shed, pull out a client box and read me the number on the side". He notes the number, dials into the server, and using WiseManager he changes the workstation definition to include the new NIC id the user read to him over the phone. And it boots! No changes to applications or user information is required. It can be that simple to recover from a catastrophic event.

Simply put, there are not many reasons for Fred to visit the site. New applications can be added. So can upgrades or modifications. Changes in users' desktop design, access to data or different sets of applications, creations of new users. All can be done remotely. Only OS/2 permits this kind of technical elegance, simplicity, and most important, the reliability to make this possible. This network was designed for availability and the function designed into it means that there will be few reasons for Fred to take active steps to manage the site and certainly, as noted, few reasons which would require his presence at the site. But his "royalty" checks come in every month. Sort of like the old IBM "time sharing" model. Customer didn't own the equipment, sometimes they owned the application. But generally they were paying for access. IBM did ok with that model.

Imagine getting beeped at midnight, turning to the lap top on the night table, dialing into the server and resetting some icons, then using FaxWorks to send a $300 invoice to the user for an off-peak emergency support incident. Then rolling over and going back to sleep. Leave the midnight runs to those supporting other products. I'm not suggesting gouging a customer. But when a call to a help desk costs $150, being able to provide the kind of service that puts a user back in business in just moments, any time of the day or night, has value. Fred deserves compensation for being smart enough not to give up on OS/2.

For Verticals, this may work, but what about the "Horizontal Market," the general business user?

Agreed, vertical applications make it simple because they are easy to replicate and tend to be the principal, if not the only application on the workstation. But general business users are also an excellent opportunity. The "casual business user" is about 80% of total business computer users. These casual users don't know much about their PC. I call them "application literate" as opposed to "PC literate". And that 80% are difficult, and therefore expensive, to support. They require an inordinate amount of the support resource in the traditional fat client PC model because they can't resolve their own problems.

Casual users also tend to disrupt the work environment by invoking the generally unproductive function I will call "peer support". In other words, they interrupt and bug their co-workers trying to find someone who can help them out. Now there are two people who are not working, more if others are "looking on". It calls to mind the axiom "You can't make something fool proof because fools are so clever". Fortunately for Fred, those 80% are excellent candidates for "managed client" systems. Fred is just the guy to manage those users as part of a complete desktop services contract.

Fred can go to accounts and tell them "You don't really want to own your IT. It costs you too much to support it and you will never have the in-house technical resource to do the job right. And any resource you expend on IT is resource taken away from servicing customers. When you get down to it, which supplies more revenue to you; managing your IT or servicing your customers?"

Part of what makes this scenario so attractive to user organizations is that they are worn out by the churn of technology upgrades. (Thank you for that, Microsoft) They would like to get off that treadmill and go back to the core business activities. That's part of what has made outsourcing so popular today.

What makes all this work is OS/2's ability to perform so reliably in a diskless RIPL environment. It is just not possible to do this with any other Intel OS today. And OS/2 is reasonably priced for this model.

But wait! What about the applications?

The chances are that most applications run by small and medium sized accounts can run without modification under OS/2. This is because so many of these apps are DOS or supported by OS/2's subset of Win32. This protects the investment made in application software, which may not have been significant, but it also protects the investment users have made in training.

The show stopper until now has been Microsoft Office. But StarOffice has come along to provide the ability to exchange documents with those users. Smart Suite Millenium is good for this, too. But StarOffice is snappy in a network and I'm suggesting diskless RIPL clients. Smart Suite is really a fat client PC product not well suited to this environment.

And lets not forget, there are still many, many excellent OS/2 applications available. Products which are superior to Windows counterparts, more functional and more reliable. In line with this, OS/2 provides a world class Java Virtual Machine. So, while it's true that applications can be an issue, it can be addressed many times. Not all, but often.

And there is a trade off. The desktop management service provides the user organization with simplicity, cost savings, a computing environment which is more reliable and available to service customers using less in-house resources. Does the user organization want to do business, focusing on revenue steams and business objectives, or do they want to use MS Office? How does that help them? This is a better argument to make as a service than as a purchase decision.

If there are requirements for other PCs, likely to be running Windows, Fred can add their management to the same contract as with the Managed Clients, connecting them on the same Warp Server. It simply means that Fred will support those Windows users with the traditional model. This won't be as efficient as the Managed Clients but maybe he can convert some more of those Windows workstations at this account later.

And what's good for the goose ...

I have focused on a consultant working with small to medium sized accounts because there are so many of them and there is a significant opportunity for many members of the OS/2 community to participate in this model. But the fact is this model works in many larger accounts as well. Client management is a challenge in any computer environment and this model goes a long way to addressing many key areas of client management. However, large organizations may have a significant IT investment to protect in products and skills. So, expect a different set of issues here.

If there is a significant number of existing OS/2 workstations, WiseManager and Managed Clients should be positioned as a management solution. In a Windows dominant environment, try focusing on the casual users. The IT organization would probably be delighted to "outsource" support for those. The casual user is the most difficult for IT to support on fat clients, but the easiest for Fred to support on Managed Clients.

Remember, Warp Server ... the same server supporting these managed clients, can also support Windows and Apple clients, as well as Novell, NT and *IX servers. So, managed clients can go where they are a good fit, all part of the same, happy network. And IBM recently demonstrated a Win32 WorkSpace on Demand client. This means some of those Windows machines may be able to participate in this model later this year.

Plus, is the Windows story so very pretty this year? Folks moving to Windows in 1999 and 2000 have a very real dollar exposure and the entire industry points this out in a different set of articles every week. Why not go to the safety of a Managed Client until Windows sorts itself out? Fred might even imply that his Managed Client is delivering on the Zero Administration for Windows promise. It will make the eventual migration to Windows more easily managed. But the reality is that once a Managed Client is in, working, and well supported, you may be hard put to find users ready to go to a less responsive, less reliable, and more expensive model.

I would add one more "positioning" remark. I have used the term Fat Client to refer to traditional PCs. That being PCs which function as PCs without a network connect, but may be connected for file and print, messaging, and so forth. I have used the term Managed Client to refer to the OS/2 diskless RIPL client, in this case using WiseManager and WiseClient. I have stayed away from "thin client". The industry tends define thin clients as computers which execute programs in the server, like Windows Terminals and Citrix WinFrame. This is not how Managed Clients operate.

Once these machines boot, they are much more like traditional PCs and NCs. Managed Clients are robust PCs running OS/2 Warp 4, not a selected subset. This is one of the reasons this model is so much more cost effective than NCs, Windows Terminals, traditional PC networks, even WorkSpace on Demand. The Managed Client, the complete desktop service that I described here, utilizes the best characteristics of each computing model. Fred can look his prospect in the eye, unflinching, and describe it as "the best of the rest". How's that for marketing weasel words?

Fred closes with, "We'll be setting up a complete computing environment for your use, directed at supporting your business objectives. Familiar applications and your current databases will be preserved. But you'll never have to buy another piece of software or hardware again if that's the way you want to go. When would you like to get out of the business of managing PCs? And, more to the point, how fast to you want to get back to servicing customers?"

And no small part of the Y2K angst should be ignored because OS/2 is Y2K ready. IBM is not mincing words on that score. But the verdict is still out on when Windows well get there.

Can OS/2 be sold as part of a service rather than as a product? Bob St. John is ready to discuss further details and counterpoints in our interactive forum, open now. Selected feedback will be posted below.

• All I need to know - Gary. (5 replies)
• One small problem - John J Urbaniak (1 reply)
• Expense vs. Capital Acquisition - Bob St. John (1 reply)
• Eating Our Own Dog Food - Kim Cheung (2 replies)
• Software License Agreements - draggook (2 replies)
• Intel Getting Into "Desktop Management" - Bob St. John (2 replies)
• How about beginning users? - jt (1 reply)