Summary: As a former IBM employee and now IBM Business Partner, Bob St. John has a clear understanding of why a new OS/2 client would make sense for IBM.

The day before I wrote this I sent an e-mail to some former colleagues at IBM NCSD (Network Computing Software Division, the division now responsible for OS/2) regarding the possibility of a new OS/2 client based on the OS/2 Warp Server for e-business; Aurora. It was triggered by the survey results in the Ziff Davis Sm@rt Reseller magazine on this topic. Considering the channel orientation of SR, I was surprised that OS/2 warranted attention. Not only that, but the results were overwhelmingly for a new OS/2 client. The key message: a new client would help me sell the new server product.

This is a message has always been a problem for IBM. IBM deserves its excellent reputation for innovative technology, for technical elegance and reliability. Couple this with IBM's presence in every other aspect of computing, mainframe, storage, networking, AS/400, RS6000, Internet, service, even financing. Amazing.

And there is IBM's ability to understand the enterprise as no other vendor can. IBM's account teams have enterprise penetration and influence that may be the company's most significant asset. But the message from the Sm@rt Reseller poll may be indecipherable by IBM.

The "Intel space" ... the computing environment once referred to as "IBM PC Compatible", aka Windows compatible or "Wintel", has always been a marketing challenge for IBM to understand. Future graduate school case studies will revolve around IBM's failures in this computing environment. IBM's reluctance to commit to a new OS/2 client is simply the most recent chapter in this case study.

While still in IBM, in May 1997, I attended a marketing meeting and demonstration of "Bluebird", the product which became WorkSpace on Demand. At that meeting, the Director of Strategy and Market Intelligence told the assembled audience "It turns out you need a client to sell the server". Everyone chuckled. I'm not sure he intended the remark to be humorous. "Well, of course you do" was the murmured response. It sounded more like "duh!" but they meant "Well, of course you do".

The issue was and continues to be that the customer -- the user -- views their computing environment as being seamless. It is their computing environment. One entity. Vendors want to divide the environment up, applications, networks, platforms, devices ... but users don't. To them it is one asset, one tool. To gain acceptance for an isolated component ... a NOS (Network Operating System) in this case, is very, very difficult. Ask Novell. Ask Microsoft. I would not suggest asking IBM at the moment. IBM appears to be in denial, again.

An aspect of this user decision making which has eluded IBM on the conscious level is the fact that many of these decisions are not simple product selections. The vendor is being selected as well. Every decision to buy Microsoft Windows NT is also a rejection of IBM. Harsh words, but often true.

IBM would argue this point. Fine. Perhaps I'm mistaken. Let's look at some marketing campaigns from 1997 and 1998. How about the "IBM sells more software than Microsoft", remember that? It omitted how much of IBM's software revenues come from Intel space and comparing that apple to Microsoft's apple.

Then there was the "IBM is the leading software vendor for applications running on Microsoft Windows NT", which proceeded the expansive "IBM suites for NT" campaign. This must have referred to leading in the sense of having greatest number of products, not leading in sales volumes. And those "IBM NT Suites", how are they doing? Making people forget BackOffice, FrontOffice, any office?

The challenge for IBM is the fact that they have superior applications and technologies, they have superior account penetration. IBM ought to own this market. But, as long as IBM cedes the basic platform to Microsoft, they also cede the environment because the environment is more homogeneous than IBM realizes. Not homogeneous in the strict product sense. It isn't that simple. It involves thinking as a user of technology to service customers. In that sense, the computing environment is not a set or products. It is one tool.

As a vendor IBM has said, in many ways, "this is not a market we choose to serve". Despite the moneys spent on development and marketing, the IBM message which is communicated is not clear or encouraging. This reduces IBM's attractiveness to these users. It actually is a market IBM wants to serve, but IBM has never mastered "walking the talk". The actions belie the words and that's what counts. Not the good intentions.

The United States steel industry made many mistakes, not just one. But a significant mistake was choosing which market it wanted to serve; ceding other markets to competitors. Those competitors were grateful to have those niches and developed into profitable enterprises until they were able to compete and win against US steel interests. It isn't that simple but, in a way, it is.

To bring this down from high orbit and return to the finite decision regarding an OS/2 client, IBM will deliver Warp Server for e-business this spring. It will be a beauty. Will anyone care? Sure. But how many more would care if there were also a positive message about the client in the material?

IBM needs to deliver the new OS/2 server with a statement of direction regarding the new client. An Aurora client would be the right complement to Merlin and WorkSpace, not to mention JavaOS. A family of IBM clients to work with and the excellent support IBM servers provide for Windows, Apple and *IX clients, and servers. This would allow marketing people, IBM and channel people, to weave a compelling story of client and server support; true cross platform support.

And this strong platform story strengthens the application story. UDB, MQSeries, Notes, and others. Only IBM is in the position to tell this story; "support for your environment the way you choose, protecting your investment in product and skills. Don't accept a vendors decision, make your own." And the IBM story can be told with superior products by superior people.

But, like the man said, it turns out you need a client to sell a server. And you need the client and the server to sell database. And you ... well, you get the picture. If IBM wants to be a force in the Intel space, it must exceed expectations of the market. Producing a new OS/2 client buys IBM some credibility in an area where it desperately needs it.

I've heard that some in IBM are saying, "It's ok not to do a client. We told people we wouldn't do a client, so how would we look if we did one now?" This reasoning brings to mind the "It hurts when I do this" and the response "Well, stop doing that". People didn't like the "No client" story, so stop telling it, or at least change the ending. Let's forget credibility and go directly to relevance, if IBM wants relevance, it must make the market forget OpenDoc, Taligent, Kaleida, CommonPoint, micro channel, and I don't have to stop there, but I will.

IBM has to stop sending the "don't use our software" message. This is, in fact, what IBM does when it fails to support OS/2. Specifically, I'm referring to the failure to support OS/2 with IBM programming tools. OS/2 has a home in the international enterprise and those accounts have in-house applications. That means they have in-house programmers. I'm not talking about shrink wrap or games. I don't want to go down that road today.

But IBM claims to listen to the enterprise account. I watched a large transportation company change from OS/2 to NT, almost 10,000 licenses, and it wasn't pretty. What drove the change was the inability of the enterprise in-house programmers to get good support for programming on OS/2. And it just isn't a large leap from "developing on NT" to "developing for NT". This is simply another lyric in same song; lose the programming tools, lose the platform, lose the applications, lose the account.

So, I want to give some simple messages to IBM Software Marketing, specifically the International Product Development Team. You will always derive greater benefit from things you do than from things you don't do. Performance against expectations is more important than performance. Offer users the opportunity to do business with you. Do not cede advantages to competitors.

My business affiliations with IBM include being an IBM Business Partner, authorized IBM PC Reseller, Software Developer Organization and IBM BESTeam member. I think I can deliver better business solutions using IBM products and technologies than working with any other vendor. But like many other IBM business partners and customers, I sure wish IBM would not make it so difficult for me.

Here is an area for improvement. IBM, please finish Aurora. You are doing great, it is an excellent product. I hope the fact that it is coming from IBM will not hold it back. You could help it by issuing the following statement. "We are continuing to evaluate delivering a workstation client based on this server kernel. If the market demand exists, we think we can manage to it and deliver a quality and exciting product later this year." Making such a statement will help the entire Software Marketing Group.

I'm not saying that folks would have to step back to avoid being trampled in the rush for the new OS/2 client. But I see the demand and I don't doubt a new client would be profitable. IBM Software Marketing and NCSD should be more than a little embarrassed by the outstanding sales of the OS/2 client in 1998. It is testimony to the excellence of their direct sales force, the product, and the reality that an OS/2 market exists.

The fact is, the market is willing to consider alternatives to Windows. This trend is growing. In my opinion, Windows has hit the high water mark. Windows does not have a good story to tell in 1999 and 2000 and this opens the door for products like, Linux, Java, and ... yes ... even OS/2. OS/2 may never regain the market acceptance it had in 1995. Remember when there were three trade publications about OS/2? But it can be much more than a jumping off point for users heading to other platforms. OS/2 can still be a destination. It is an excellent product.

Can IBM improve the Intel space revenues? Of course. Will a new OS/2 client do that? Of course not. But a journey of a thousand miles begins by reclaiming some credibility in this market.