Behind Big Iron's Bad Quarter

Although Big Iron had a bad quarter, the long-term indicators -- bolstered by several mainframe product cycles' worth of data -- suggest System z is just tracking through its life cycle.

As expected, IBM Corp.'s mainframe fortunes took a tumble in the last three months of 2011, with overall sales dropping by almost one-third, or 31 percent.

Should Big Iron watchers be worried?

Probably not. Quarter-to-quarter volatility isn't a useful predictive tool -- it's the longer-term trends that matter, not what appears to be trending – and IBM itself predicted a downturn in mainframe fortunes. Context also matters -- as CFO Mark Loughridge stressed in his remarks last month to financial analysts, System z's performance in Q4 of 2011 was down relative to its performance in Q4 of 2010, which itself was the first full quarter during which Big Blue's then-brand-new zEnterprise 196 mainframe was shipping.

There was a lot of pent-up demand for zEnterprise. In fact, sales of IBM's z10 mainframe had seemingly evaporated in expectation of zEnterprise, with the result that Big Blue actually moved up its announcement of zEnterprise by a few weeks.

"[W]ith the exception of mainframe, which was coming off of the biggest quarter in its history last year, each of our 16 brands across the company gained or held share," said Loughridge, specifically highlighting the strong performance of IBM's Unix business.

"System z revenue declined 31 percent as we wrapped on the successful launch of our zEnterprise 196 last year. MIPS declined 4 percent this quarter. System z gross profit margin was up, reflecting a higher proportion of microcode upgrades, which is typical at this point of the product cycle. POWER grew 6 percent. This is the 15th consecutive quarter of share gains in POWER."

IBM's mainframe drop-off had a noticeable impact on the overall performance of its Systems and Technology Group (STG), which posted sales of $5.8 billion -- an eight percent decline year-over-year. "This performance reflects very strong growth last year, driven by mainframe growth of almost 70 percent. Gross profit margin was down 3 points year-to-year, with a significant impact from product mix," Loughridge confirmed.

IBM had a good run with zEnterprise. Building on its record-breaking performance in Q4 of 2010, Big Blue's next-gen mainframe was a powerhouse through most of 2011. In the first three months of last year, for example, Big Blue sold more than $1 billion worth of Big Iron hardware. On a year-over-year basis, System z's performance in the first three months of 2011 was up by almost half (41 percent) from the same period in 2010. That's quite a performance, but in Q2, IBM did even better, generating $1.2 billion in mainframe sales according to market watcher International Data Corp. (IDC), which was a 60 percent increase over Big Iron's Q2 performance in 2010. (True, Q2 of 2010 was one of the worst quarters in mainframe history, but even so, over $1 billion in mainframe-related revenue is impressive.) Q2 also marked the fourth consecutive quarter in which Big Blue's mainframe group generated $1 billion or more in revenue. It would also be the last such quarter.

Big Iron didn't -- and really couldn't -- sustain this kind of performance through the third quarter of 2011, although this, too, was expected. In fact, zEnterprise's $970 million performance in Q3 -- which saw a 5 percent decline in sales over and against a period in which the then-new zEnterprise first started shipping -- was more or less in line with IBM's own guidance.

In his Q3 2011 remarks to financial analysts, Loughridge conceded that the pent-up demand that helped fuel four consecutive $1 billion-plus quarters for zEnterprise probably was spent. More to the point, zEnterprise has been shipping for almost two years (19 months); at this point, it's at least halfway into its life: Big Blue typically revises its mainframe CMOS every 24 to 30 months. If history is any indication, we might expect demand for zEnterprise hardware to start dropping off even more precipitously by the third and fourth quarters of next year.

IBM can try to offset slackening demand by slashing prices, as it did in the first few months of 2010, during which, not coincidentally, System z posted some of the worst numbers in its history. There's some evidence that Big Blue slashed prices in Q4 of last year, perhaps in expectation of -- or as a response to -- slackening demand.

The short-term indication -- based chiefly on its poor performance in Q4 of 2011, but also anticipating softer sales in 2012 as a whole -- is that System z is once again imperiled. The long-term indication -- a conclusion that's bolstered by several mainframe product cycles' worth of data -- is that System z is just tracking through its life cycle. The caveat here is that it is a distinctive life cycle, inasmuch as the mainframe -- in spite of IBM's efforts to drive down costs -- is still perceived as a premium (or in some cases prohibitively costly) proposition. This means that there's typically plenty of pent-up demand for new mainframe CMOs, thanks to a number of factors, not the least of which is IBM's much-ballyhooed mainframe technology dividend.

For this reason, purchasing mainframe hardware isn't like purchasing Intel-based or (even) Unix-based equipment: mainframe shops are more likely than are their open-systems counterparts to put off purchasing gear, at least with a next-gen offering waiting in the wings. In the present case, the pent-up demand that's a product of deferred purchasing seems to have exhausted itself, and mainframe sales will likely slow even more in the next few quarters.

However, to conclude from this that the mainframe as a platform is -- once again -- imperiled is to ignore the data. Or, as veteran mainframe watcher Alan Radding puts it, it's to reprise an embarrassing mistake. "[J]ust in case you thought the zEnterprise is going away anytime soon, don't worry. The trends are headed in the zEnterprise's favor. IBM added 62 mainframe clients in 2010, 76 new mainframe clients in 2011, and expects to hit 100 or more in 2012," Radding wrote on his Dancing Dinosaur blog. "Remember all the pundits over the years who predicted that the mainframe was a dinosaur heading to extinction? Don't bet against the zEnterprise."

comments powered by Disqus