Reader Feedback: Declare War on Wintel
Published: February 20, 2006
by Brian Kelly
I like Elaine Lennox, the new vice president of iSeries marketing. I have never met her, but based on her interview in this newsletter, which was published last week, she sounds real. I hope all System i5 folks let her know what the iSeries is all about.
Unfortunately, those potential new System i5 accounts that are now running Wintel, who never considered a System i5 for whatever reason, cannot tell Lennox what is needed and that is who she needs to hear from. The System i5 must look modern and it does not and IBM needs to do something for the developers that keeps the system afloat. More and more is needed and I have talked about that for years, but many of the things we need do not appear to be coming. I hope Lennox can change that.
Lennox stated the crux of the problem in her interview:
"…what we are trying to do is take the iSeries back into the mainstream, into being a core growth platform as part of IBM's Systems portfolio. We want to get new customers and re-energize it in the marketplace."
The System i5, formerly the iSeries, formerly the AS/400, and formerly the System/38 has always been the most unique and the most technologically advanced system from IBM. It was always way ahead of the pack in its technology. The 1969 System/3, the 1977 System/34 and the 1983 System/36 were the IBM systems that were best known for simplicity in implementation and pure ease of use. It helps to remember that in June 1988, with the introduction of the AS/400, IBM chose to combine the technical prowess of its System/38 with the loveable qualities of its System/36.
Unfortunately for IBM's System/36 customers and for IBM itself, the System/36 part of the combo was cumbersome to implement and not simple like the System/36. Worse than that, it performed very poorly compared with the real System/36. As an IBM Systems Engineer during this period, I witnessed first hand the negative feelings that System/36 clients began to harbor for IBM and the new AS/400--especially those who spent lots of money and whose machines ran poorly.
Between 1988 and 1991, System/36 users would do anything to not have to move to the AS/400. As long as IBM offered upgrades, they would buy as much as they could to keep from switching. Then, when there were no adequate upgrades, they would search for more powerful used machines. They would move from the little 5363 to the bigger 5362 to the biggest 5360 models. The 5360 clients would try to get a second machine. Some, in desperation, switched to System/36 emulation facilities that were offered on Unix platforms, including IBM's own RS/6000 AIX servers.
Though the company cared, IBM did not have to care that much because the AS/400 was very successful picking off DEC, Data General, Wang, and Hewlett-Packard proprietary systems. The AS/400 helped end the minicomputer bleed for IBM and the AS/400 showed great promise--ironically, just a few years before mainframe downsizing came into vogue and people started looking for new platforms on which to run COBOL applications. I had just finished MVS/DB2 school at the time. Just prior to the AS/400 launch, it did not appear to be a safe bet to be a midrange SE with the Fort Knox mentality still prevalent in the company. (Endicott mini-mainframes were to take over Rochester small systems.) The AS/400 and its rapid success made it safe to be an IBM midrange SE again.
For about three years and maybe four, the AS/400 ruled the day. It was the right package at the right time to literally eliminate the other proprietary minicomputer vendors. HP is the only real survivor from this pack and if HP did not have such a great peripheral line and had to live from its minicomputer revenues, HP might not be around today. That's how good the AS/400 was for its day. Unfortunately, IBM stopped investing the right dollars in the platform.
In 1992 and 1993, the COMMON crowd would often complain at its famous Soundoffs that RPG or CL or something else AS/400-specific was not being enhanced. IBM, it seemed, was not offering much new to its constituency during this period and it seemed that the company had begun to position the box to be sold only to existing users.
For example, client/server computing had already been made popular by Novell. Microsoft cared enough about client/server that it had already failed with two products that were aimed at dethroning Novell as client/server king. Yet, IBM was not concerned in the least about Microsoft nor did it think Novell was a threat. Bill Gates was in the process of building Windows NT at the time, with the purpose of stealing Novell's client/server business and more. Still, IBM chose not to invest in client/server. The rationale as explained to those of us in the field was that IBM already had that technology covered with its PC and PC server line as well as the RS/6000 line, the company's Unix box that was launched in 1990. So, IBM consciously decided that it did not need to invest to provide the server side of a quickly emerging technology, client/server, on its prized platforms. AS/400 shops were forced to buy PC servers to do simple file and print serving.
In the early 1990s, another new technology was rapidly coming into its own. In 1990, for example, the whole world came online with world.std.com firing up as the first commercial provider of Internet dial-up access. You know the rest of the Internet success story. In November 1990, the first World Wide Web site came online, and in 1991, Tim Berners-Lee introduced the CERN World Wide Web server to the world. By 1992, there were over 1 million hosts on the Internet. Despite Al Gore's promise of the Information Superhighway, and IBM's major corporate commitment to build the Internet backbone, none of these hosts were AS/400s.
Just as with client/server, IBM chose to reserve these new technologies for its non-proprietary architectures--PCs and RS/6000 Unix boxes. In fact, the AS/400 TCP/IP stack in the early 1990s cost more than the RPG compiler and when somebody actually bought it and used it, the system ran like molasses. This overpriced dog existed only so IBM could bid on contracts requiring TCP/IP. Big Blue was still counting on SNA for its AS/400 and mainframe platforms. Nobody was thinking Internet until Lou Gerstner showed up on April Fool's Day, 1993.
Gerstner noted quickly that the Internet had the greatest potential for IBM's servers because the Internet required powerful servers and IBM made powerful servers. Already several years behind, it took until 1995 to get a Web server operational on the AS/400 and it took even longer to get all the pieces going on the platform. Unix servers had this stuff right from the get go. Microsoft fought the Internet at first, but once Bill Gates saw his mistake, he wasted no time gearing up his juggernaut to dominate the Internet server world, and he turned the company around on a dime in 1995. While IBM was building its Internet stack and applications, the company also added some technology that had been missing--the server part of client/server stack. But IBM stopped short of permitting its developers to actually engage in creating applications for this new environment in a fashion to which they had become accustomed.
It is still hard for me to grasp why it took a guy with a background in cookies and credit cards to change the future of the AS/400 and the mainframe by insisting these proprietary IBM systems be retrofitted with Internet technology so as to be competitive in the e-business world. Yet it is true. When IBM digs in on a thought it often stays there lots longer than it should. The PC and the PC server, even those with IBM on the front cover, were not IBM unique products. The IBM PC was piece parts from many vendors. IBM was not even the leader in PCs in the 1990s. The clones had already won. Similarly, the RS/6000 had a processor built for a workstation computer, not a server. Early RS/6000 offerings did not have real scalability (and neither did anyone else's Unix servers, to be fair). Yet the AS/400 role in all of this was kept at a minimum by IBM decree.
In retrospect, it is easy to find it odd that IBM could have decided that the PC and the RS/6000 were all that it needed to be successful with client/server and the Internet. Was it to avoid development costs or because the company actually believed it was well armed for the Internet battle? From having worked for IBM right up until this time, my bet is that it was both. Before Lou Gerstner, IBM was fully prepared to keep its most powerful and proprietary machines (more profit) on the bench.
Because of all those years on the bench, both the AS/400 and the mainframe have had a tough time catching up with the Unix and the Windows crowd from a functional standpoint. See for yourself. The front door of each of these systems still is still adorned with a green-screen interface. No wonder those folks, who don't know any better, think the AS/400 and the mainframe are both legacy machines. Why would anybody who sees a PC running 5250 emulation or, worse yet, a 5250-type or 3270-type terminal connected to one of these monsters, conclude that the system consisted of anything modern or worthwhile? A cliché such as a picture is worth a thousand words is certainly trite, but it does make my point. One might suggest, therefore, that the full Gerstner directive on the AS/400 has yet to break through that 1994 mentality--even though Gerstner has been gone for four years--and the AS/400 is now on its second name change.
While all this was happening with the AS/400 in the early 1990s, the System/36 shops were still disenfranchised. In our local IBM branch office, occasionally I would see a System/36 shop that was against the wall performance-wise. They would spend the dollars and move to the AS/400. But most just sat still. In 1994, IBM decided to sell some AS/400s to System/36 customers. To do this, the company re-introduced the System/36. Big Blue called this new small system, the AS/400 Advanced/36. I am not sure whether Gerstner suggested this or it was Rochester trying to reel in the lost sheep,
Douglas Adams once said, "If it looks like a duck, and quacks like a duck, we have to at least consider the possibility that we have a small aquatic bird of the family anatidae on our hands." This new machine looked like an AS/400, but it quacked like a System/36. That quacking was enough for most of the AS/400 holdouts to flock to the system in "droves." And flock they did. IBM's announcement of the Advanced 236 and then the Advanced 436 models were, to some, viewed as its way of apologizing for closing the System/36 door six years earlier. IBM came out with a new System/36 in 1994 that was as welcomed by its to-be constituents as the System/36 itself was in 1983--on its original day of announcement. For giving them what they wanted, Big Blue's customers rewarded IBM with many orders for this new "AS/400." What an interesting marketing idea: Give your customers what they want and they will buy it.
So once the System/36 wild ducks came back to Silver Lake (the AS/400's development code name from 1987), the customers who wanted ease of use (System/36) and the customers that wanted high tech (System/38) that was easier than Unix, Windows, and mainframes all were essentially able to use the same hardware platform, the IBM AS/400. Somewhere along the way IBM forgot about getting any more new accounts.
In 1995, RISC and 64-bits came to the AS/400 in one of the most unheralded announcements of all time. It was not until 2001 that the same capabilities came to the IBM mainframe. That, too, was quiet in terms of press coverage because after all, it was six years late. However, as Intel struggled with its Itanium--or as those with a good sense of humor have dubbed it, the "Itanic"--the press (present company excluded of course) pretended that it had never heard of 64-bit IBM Power processors and it made the Intel Itanic appear as if it were going to save the world. The IBM AS/400 had already saved the world in 1995, but the press had obviously misfiled IBM's enticing press release.
It seems to me that once IBM started the CISC-to-RISC migration, the new account machine all but closed down. The $2,000,000-gross-sales furniture store down the road not only did not hear the CISC-to-RISC marketing messages for the last five years of the 1990s, it wouldn't have understood what it had to do with them if they had heard.. What kind of mumbo jumbo is that?
Gerstner had surely saved the big and burly IBM from sinking, but he also got rid of the IBM sales staff. Selling a push system with 64-bits, without your own experienced field sales team would be a chore for any company. While I was with IBM, a few business partners sold a few new systems by themselves--but not often. In bigger cities, I would suspect there were better results, but I really do not know. This is not a jab at business partners. Because of the political climate within IBM in the late 1980s and early 1990s, many local business partners were rewarded with commissions for systems that were sold by the local IBM branch new-account marketing representatives. If the reps came in with an order and it was not sold by a business partner, it was almost worse than no sale. IBM local management would fry them. It did not take long for IBM reps to understand that meant that each sale had to have a business partner assigned even if they provided no sales assistance.
IBM, of course, paid commissions to the business partners because, obviously, they had sold the system. The marketing rep also got paid. Local management got paid. They called it win-win-win. In many ways, that is the same as a corporate welfare state in which "nobody pays." Yet, somehow the taxpayer pays. In the business partner scenario, IBM paid for something it did not receive. Worse than that, IBM corporate and IBM local management were happiest when sales were made by the business partners. In the branch office, however, the locals knew that IBM was paying for services not received, but even worse than that, the sales results indicated that the switch to business partners from direct sales had already been a success. So, as expected, by 1995, IBM had laid off almost all of its sales force. After all, business partners were bringing in the business. When IBM digs in on a thought it often stays there lots longer than it should. That behavior almost bankrupted IBM and it cost John Akers, Gerstner's predecessor, his job.
Eventually, without an internal IBM sales team, business partners were in many cases awarded exclusive control of installed AS/400 accounts by IBM. Whether they needed to or not, many of these shops were ready to buy a system every four or five years. If the lease was up, that was a great opportunity for a business partner to spend the time to get an upgrade or, more than likely, an order for the newest AS/400 model. But these folks were not being paid the kind of money that it took to sell a new account when the alternative solution in all cases was about half price.
By 2000, there were tons of new startup businesses all over the world. They were popping up and becoming successful. At the same time, the normal mix of mom and pop businesses were growing up to midrange size. None of these successful entities needed IBM nor its business partners to tell them to automate their operations. They were not stupid. By 2000, everybody knew that you either automate or fail. And for these newly successful growing enterprises, Microsoft, along with Dell and HP, had enough advertising underway that these automation-wannabes had no problem finding somebody to sell them the exact technology they "needed." Well, OK, not exactly! But they thought it was what they needed because IBM was telling them nothing.
To make matters worse, when IBM checked the numbers and noticed what was happening, it realized that it had no real push, no new-account sales force to turn it around. It was not financially rewarding for a business partner to sell just five or ten new accounts in a year because IBM gave them no incentive to do so--as it had once given its own sales team. IBM realized that it had better start attracting customers to its proprietary SMB platform using a different marketing formula. If there is nobody pushing a product, then somebody has to pull.
However, just in the nick of time after AS/400 loyalists were promised that IBM was about to begin to advertise the AS/400--even on TV--Big Blue, and I mean Big Blue put the kibosh to the deal. When the marketers in IBM Rochester finally realized they had prematurely pulled the plug on AS/400 new accounts marketing by externalizing the sales team before the partners were really tuned in, they had to do something. They knew that short of rehiring some sales people, they had to change from a push strategy to a pull strategy. They knew that within the corporate IBM culture, they would not be allowed to declare a mistake and correct it. This would not be permitted since among other factors, "IBM does not make mistakes!"
So it was time to get some advertising messages out quickly. With nobody selling new accounts and with an IBM that would not permit itself to suggest what was needed if it was different, Rochester was in a pickle. Somehow, by hook or crook, it had to get a message out to new computer prospects that there really is a reason to look to IBM for a new business solution. Advertising was the key. But, with IBM's big year 2000 plans, no eServer iSeries 400 advertising was permitted.
Too bad Rochester! In 2000, IBM was taking control of all its server divisions--from PC to Unix to AS/400 to mainframe. Thankfully--and I am kidding when I say that--once the sales force had been eliminated, IBM had a group of folks who had not been displaced, yet that had little to do. These were the former "phone mail police."
The Parable of the Phone Mail Police
Let me tell you a little story. For decades, while the world spoke otherwise, IBM had refused to call voice mail voice mail. It was phone mail. And, these KGB types had been able to intimidate the local IBM teams into changing their phone mail greetings for the smallest reason. Valid reasons, for example, were when employees went on a 15-minute break or go to lunch or go to the bathroom.
This group of marketers (former phone mail police) had gained so much influence within the company that they decided to stretch their empowerment. They suggested to IBM's top management team that the reason IBM was not as successful selling systems as it should have been was that the buying public was confused because IBM had four server products and that was too much for the public to sort out.
Because Kellogg, the cereal king, had 49 products that all offered subtle advantages over the others and GM had 73, these two organizations were not cited in the "police" study. The phone mail police cited that it would be better to call all of the servers the same name and let the consumer pick their favorite letter to decide which server division should get the business. So, because Gerstner loved the letter "e" (presumably for "electronic") and IBM Rochester could not deny it, they proposed that the letter "e" be placed in front of the term "server" to make a brand name: eServer. To placate Gerstner in 1997, when IBM coined the term "e-business," the Rochester team had placed a little letter "e" at the end of the term AS/400--as in AS/400e. So, there was already a Rochester affinity towards the "e."
The "police" astutely argued that if IBM were to go to just one product, the nasty press that for years had claimed that IBM was plagued by having too diverse a computer system product line would be silenced. Since the "police" were the only ones with guns permitted in the room, their motion carried and management approved their suggestions. All IBM systems would henceforth be known as eServers.
Then there was the letter drawing. All letters of the alphabet were placed in a big bowl. The mainframe division got to pick first, then in order of importance, the PC division picked, and then the Unix division. After the drawing, the letter "i" was found on the floor. Not having enough budget to ask anybody from Rochester to fly in to attend the drawing, it appeared as divine intervention that Rochester's letter fell to the ground. So let it be written. So let it be law. Rochester would be "i."
There was one member of the phone mail police who believed that one letter was not enough. He liked the idea of eServer, but felt that eServer i for example didn't give enough to be liked. He remembered a machine in his supermarket that once made sure the freezers were all running with less power. He said it had an IBM on it and the word Series with a number. To make it a new idea after all, he suggested that the letter go before the word Series. And, thus, after all these years, we have the true derivation of that well-loved term, "eServer iSeries."
The rest of the phone mail police immediately accepted the extra letters. They recalled other brilliant managers having said that you should not pay any more for something than you should. Since the extra six letters cost nothing, they approved the lengthened term with the "i" in front. Since z, x, and p were the other letters pulled, you already know the rest of the story.
But, what about that 400 thing, as in eServer iSeries 400? After the phone mail police had done their work and it was accepted at the highest levels of the corporation, a telegram that had been delayed due to a snow storm in Rochester finally arrived. It read as follows:
Hello Phone Mail Police --- Stop
Rochester Here --- Stop
Our thoughts on new name --- Stop
Must say 400 since that is our brand name --- Stop
Ga3477alkfh7734aj3552^^#891625 --- Stop
So, the phone mail police, rejoicing in the new names for all IBM systems, were in a good mood. They agreed to the addition of the numbers 400 to their choice and thus, we have the moniker eServer iSeries 400.
Eventually IBM began using voice mail and stopped using phone mail. Thus the phone mail police were laid off and management was able to run the company without these tyrants.
Just remember, all parables have an element or two of truth in them. The net effect of the eServer rebranding is that for five more years after 2000, there was no advertising for any eServer that dared suggest it would be better than any other eServer. Unless, of course, Intel--with its big bags of co-marketing dollars--was paying for that ad. And since Intel paid for a lot of ads during this period, the words iSeries and 400 were nary heard again.
So, for five years, those AS/400 prospects who had been pulled in for some reason or another over the years but had not been ready to buy, never heard of the AS/400 again because it was gone. Even though it was the only thing that could give the AS/400 product a boost since there was no push and there was no pull strategy permitted, the brand itself had to go. A brand recognition strategy could not be implemented because the brand had been deleted.
In 2005, IBM placed some folks in the iSeries management seat who really want to make the platform grow. Mark Shearer, the new general manager, and Peter Bingaman, the former vice president of iSeries marketing, were embraced by this love-starved community. Bingaman left IBM at the end of 2005, and Lennox now has his job.
". . . what we are trying to do is take the iSeries back into the mainstream, into being a core growth platform as part of IBM's Systems portfolio. We want to get new customers and re-energize it in the marketplace."
That's what she said, and that is great news.
Nobody needs to say that IBM has neglected the AS/400 in its marketing. The reason that I think IBM has noticed the AS/400 in the past year is that revenue growth from services and software have come to be a challenge. Therefore, all revenue is now important to IBM. Maybe I am wrong. It's a guess.
I am glad that Lennox requested this job and I hope that she enjoys a challenge. IBM is her challenge as is the Wintel community. IBM has not let anybody learn that an AS/400 heritage machine is different from all of IBM's other systems. Somehow, the Intel community has gained millions of new accounts and many of them can actually afford today's System i5, but they have never heard about it.
It's not that there are no new accounts to be had. There are many. If the System i5 does not have a functional, almost give-away price point as does Wintel machines, it will not get in at the bottom when somebody is looking. In my own little companies, I cannot afford a System i5 of any size. Here is a current problem list for getting new accounts:
- Nobody is selling new System i5 accounts for IBM.
- Nobody has been selling new iSeries accounts for many years.
- There is no longer OS/400 platform inertia out there.
- The ISVs are quick to change horses. Many ISVs have been wooed by Microsoft/Intel, so they have a half-price Microsoft offering that looks modern and sells better than their System i5 offering, and apparently that's OK, even if IBM invites them to the table.
- Many of these same ISVs don't care about IBM. They care only about making a buck on Wintel.
- My neighbors do not know that IBM sells small business computers and each of them have small proprietary businesses.
So, it is really this simple: IBM must declare ware on Intel-based systems.
Trying to sell new accounts and not being willing to take on Intel is like winning a scholarship but having to maintain a 2.5 average. To know you are striving for excellence, you have to strive for an A. The "A" for IBM will come only when the Intel-based system is the target.