Why Unix Experience Matters for System i5 Sales
July 10, 2006 Timothy Prickett Morgan
A few weeks ago, when I wrote up the announcement that IBM had made to appoint Peter Small as the new director of business partner and independent software vendor channel sales for the System i platform, I said that I hoped to speak to Small as soon as he got back from vacation. We did that just prior to the July 4th holiday, and he gave me some insight into IBM’s difficulties and triumphs in the Unix server market and how that experience will help pump up the System i product line.
Small has been active in the sales organization for IBM’s pSeries and, now, System p, division for the past seven years, and he is rightfully proud of the part he played in moving Big Blue from a distant third in the Unix server market–well behind Sun Microsystems and Hewlett-Packard–to the market leader in terms of technology and one of the three horses in a three-way, neck-and-neck, race for the Unix prize with Sun and HP. IBM’s gains are remarkable, and I have credited its gains (time and again) to the superior technology that IBM delivered with the Power4 and Power4+ servers and its very aggressive tactics when it came to pricing, especially when Sun and HP were on the ropes in 2001 through 2003 because their respective Unix boxes were late and the ones they sold were overpriced compared to what IBM delivered. But, according to Small, IBM was making it look easier than it was, and there is more to the Big Blue Unix story than meets the eye. There was a very big solutions play going on behind the scenes, as it turns out, and this is where Small (and Bill Donahue, who is Small’s boss as vice president of worldwide sales for the System i line) plan to leverage their Unix marketing and sales skills.
Time and again, I have said that IBM has to convert the iSeries and System i from a “push” product to a “pull” product, in terms of marketing. In the early days of the AS/400, IBM didn’t sell the AS/400 so much as take orders for it. Yeah, to be fair, you always have to educate customers about new features, and the first-generation AS/400s from 1988 and 1989 were disappointing to many System/36 customers, but they provided very affordable and powerful computing that was absolutely competitive with the Digital VAX and HP 3000 platforms of the time that, with IBM, made up the Big Four of the minicomputer market. (IBM was actually two of the Big Four, because it sold a relatively large number of baby mainframes such as the entry 937X and midrange 43XX lines.) To a large extent, the AS/400 was a pull product, particularly because the OS/400 platform–thanks to the vast number of software houses that had created applications in the RPG II and RPG III programming languages for the respective System/36 and System/38 server lines–was sold predominantly as an application server. Hence, the name: Application System/400. There were thousands of ISVs who had code ready to run, either natively or in emulation mode, on these boxes, and they rode up the wave as IBM delivered a box that smelled like the System/36 but was more powerful than the System/38, and therefore could span a wider range of businesses. The pull was due to good technology and that big base of applications.
It is important to remember that back in the early 1990s, the AS/400 was absolutely offering the best price/performance in the midrange for online transaction processing. IBM was the first vendor to adopt the TPC-C benchmark test and when it did, it ran it on the AS/400. IBM hasn’t run a TPC-C test in years on the iSeries or the System i5.
Somewhere around the mid-1990s, when Unix was the dominant server platform and Windows was just catching on, the AS/400 went from being a pull product to being a push product. Now, instead of taking orders for AS/400, IBM and its software vendor and reseller partners had to start making the case for why the AS/400 was worth the premium that IBM was charging. This is when the “total cost of ownership” and “ease of use” arguments started. What used to be an aspect of the AS/400 that customer got for buying the system in effect became a priced–but unspecified–feature. And, I would argue, over time, it has come to represent the largest single item in a latter generation AS/400 or an iSeries or System i5 server. (Some day, for fun, I will work out how much money this actually is, and then figure out if it is worth it by reckoning how many person-hours of time OS/400 saves people each year in system and database administration and programming tasks. In doing that I would, once again, be doing IBM’s work for it.)
So here were are, many years later, and IBM gets more than two times the revenue each from its System p and System x server lines than it derives from the System i. Thanks in large part to Windows being the dominant server platform in terms of annual sales–and by far the highest-volume server product in the history of computing–the System x product is a pull product. As long as IBM delivers innovation in packaging and a mix of Xeon and Opteron processors, it will get its share of the X64 server market just because it is IBM and it has one of the largest channels in the world. In terms of Unix, the situation is getting a bit tougher for IBM as Sun and HP have gotten their acts together and are delivering better price/performance. This will be particularly true by August, when dual-core “Montecito” Itaniums and dual-core “Santa Rosa” Opteron processors come to market and are peddled, respectively, by HP and Sun in their Unix boxes. But the Power5+ processor from IBM is no slouch, and Big Blue will get its share of the market, which is about one third.
IBM didn’t always have it easy in the Unix market, according to Small. “Even though Unix was a pull, AIX and pSeries were not,” he explained in an interview. “We had to figure out a way to make this a pull, even though Unix customers had already by and large chosen Sun or HP. We had a perceived ‘me too’ marketing effort, and there was a lot of feeds and speeds. Because of this, we systematically changed the sales force.”
To Small’s way of thinking, when a vendor is relying on feeds and speeds, it is pushing. And it is very difficult to push into the Solaris and HP-UX Unix server markets. “We were selling through the–I don’t even know the right word to describe this–religious zealots behind Solaris and HP-UX,” said Small. This will sound familiar to the OS/400 community, and thankfully, IBM is selling the System i into OS/400 shops that usually have a strong OS/400 culture even if the Unix, Windows, and Linux platforms also have their own zealots in their shops, too. Small says that the hardest thing was not talking to MIS managers and CFOs about the benefits of AIX and the Power4 servers, but the “illogical fights” that the salespeople and reseller channel got into with the system administrators and database administrators who liked Solaris and HP-UX and fought to not have it removed from their shops.
To combat this, Bill Zeitler, general manager of the Systems and Technology Group (which was back then known as Server Group), shifted the gears of the pSeries team and started training and qualifying sales people and business partners not just on the pSeries hardware and AIX software, but on software stacks aimed at particular industries. This was not necessarily a deep training on the application software–maybe a few days of intense briefings on software–but it gave the salespeople and partners something else to talk about, and another entry point into the conversation that would not set off the Solaris and HP-UX zealots immediately.
Interestingly, Small says that in running pSeries sales back here in the Americas region, he had to contend with another sticky problem with the ISVs. In many cases, the raw performance of the Power4 and Power5 processors allowed end users to reduce the number of processor cores in their machines by a factor of three to one. ISVs, which often make their money and charge for maintenance based on processor cores, were not exactly thrilled with the Power4 and Power5 machines. IBM had to work hand in hand with ISVs and customers to explain that by cutting down on the hardware expenses and initial licensing expenses on a collection of application software would allow the customer to spend more money on bringing in new applications, adding users, and otherwise spending more money on software. “This was a totally different approach, not just bringing down the hardware costs by 50 percent,” Small said. “We really did start going to market with the ISVs.”
Having only been on the job a few days when I talked to him, Small understandably did not have any precise plans for how to pump up System i5 sales through the ISV and reseller channels. But getting back to basics with the channel and emphasizing that the System i5 platform can be more profitable seems to be part of his plan. (Small’s predecessor in the job, Joyce Bordash, said that she would be shifting gears in 2006 from shoring up the platform’s ecosystem through the iSeries Initiative for Innovation to making sales; soon thereafter, she moved on to take a sales position in IBM’s Systems and Technology Group, which is why Small now has her job.)
The intense competition in the server market in the late 1990s and early 2000s really put the squeeze on hardware resellers, who had to ride down the price/performance curves while at the same selling more aggressively against intense competition and, in the early 2000s, decreasing demand. IBM brought the right technology to market at the right time, and could slash prices on the Power4-based machines and kick HP and Sun a bruising in the market because it has the iSeries profits to counterbalance the 45 to 50 percent discounting that was rampant in competitive Unix accounts. Small says that the solution providers today are feeling the same squeeze that the hardware resellers were feeling five or six years ago.
“A lot of the software vendors are struggling,” Small said. “There’s a thirst and a hunger among the ISVs, and they are all trying to figure out how to make margins when it is a razor-thin knife fight.” Small explained that he believes that ISVs understand that they need to profit so they can invest in real innovation, which in turn leads to more sales and then, eventually, more profits.
ISVs go into the Windows application market because they feel they have to in order to survive, but in a Catch 22, in doing so, they make it harder for themselves to survive because the only vendor that seems to make money in the Windows market is–you guessed it–Microsoft. “I would rather have 60, 70, or 80 percent of a market that delivers value and profit,” said Small. “Vendors are looking for something where they can innovate, not just participate. The Windows market is very difficult.”
One last thing. Small, like other recent hires into the System i division, chose this job, and that is obviously another factor that helps the platform. (Imagine, if you will, a sales person who was told they had to do a job, instead of seeing how they might do it and then choosing to give it a try.) “My heart is here. These are different platforms,” Small explained, referring to the System i and System p, “but they are the same customers.” He said that the secret to the success of the System p was that IBM identified pockets of customers in specific geographies and industries who would be amenable to IBM’s AIX/Power server sales pitch, and that this will be the saving grace of the System i line as well. “We sometimes get distracted by the big moves in the IT industry, but System p didn’t and System i doesn’t.”
As for my recurring suggestion that IBM actually reward those ISVs and BPs that push the OS/400 platform and lead with it–as opposed to just offering it alongside Windows, Unix, or Linux platforms in an agnostic fashion–Small doesn’t think this is a good idea. “IBM is all about open standards and choice. I want to get to the ISVs, and I want it to be a natural selection, not a forced selection.”