My history with IBM goes way back — all the way, actually, to my conception. You see, my dad was an IBM systems engineer. Growing up as the son of an IBMer had its plusses (e.g., writing my first program as a second grader — using Business Basic on a 360) and its minuses (e.g., "having" to work as the batboy on the Denver IBM softball team), but one thing was for sure: I grew up respecting IBM and, frankly, a bit in awe of the amazing machines that it built.

That said, the IBM for whom I tracked down foul balls and retrieved bats is most definitely not the IBM of today. Today's IBM primarily measures itself on the number of bodies that it places on consulting projects, not on the number of boxes that leave its manufacturing facilities. In fact, over the course of the past 40-50 years, IBM has changed dramatically. I realize that, as a reader of this magazine, you have some level of allegiance to the System i/iSeries/AS/400/S/3x platforms, so I know that you also share an affinity for a part of IBM that traces itself back to IBM's roots: building amazing machines. However, most people who think of IBM today probably see commercials with boxes draped in red cloth ("What's your big idea?") or a drive-through customer service stand ("What makes you special?"), in which IBM is portrayed as a problem-solving, approachable giant that will bring massive benefit to your company, big or small.

So, which story is true? The one in which IBM builds the finest business computer known to the human race (the System i — or is that the Power Systems and i? I'm frankly not sure!), or the one in which the company is poised to help your business capitalize on your big ideas or amplify the things that make your business special? It depends on your viewpoint. Both stories are true, but IBM's future has less to do with "machines" and a lot more to do with "international business."

Before I go down the rabbit hole and start examining IBM's business — and the relative part that the System i represents — I need to state something: This article looks at IBM from a financial viewpoint and is not intended to offer any guidance related to the applicability of the System i (or any other IBM machine, for that matter) to run your business. Think of this article as an examination that an analyst who knows little about IBM and is considering investing in the company's stock might make. My intent is to be objective and let the facts speak for themselves. (For an article that discusses ways that you can support the System i, see "Become a System i-vangelist: Promote What You Love," page 25.) In addition, the original version of this article left my word processor about three months before IBM's announcement of the Power System and the "i" operating system. Consequently, this article still has some references to a separation between System i and System p, but I believe it goes a long way toward explaining IBM's motivation for creating the Power System platform in the first place — and IBM's ongoing, continuous process of homogenizing its hardware platforms.

I've Been Worrying About IBM for a Long Time

In the spring of 2000, I gave the keynote speech at COMMON in San Diego. During that presentation, I had a chance to break down IBM's financial statements and, for the first time in IBM history (if I remember correctly), the year ending 1998 showed that hardware was not a majority of the company's revenue. Hardware accounted for about $35 billion out of a total of about $71 billion of income on operations, or about 49 percent. That means that the combined revenue for global services (consulting) and software represented about 51 percent of IBM's total revenue. Even more telling than the revenue numbers was the gross margin (i.e., revenue minus cost of sales, before subtracting sales, general, administrative, depreciation, amortization, and tax expenses). IBM earned about $11 billion on hardware during 1998 and almost $18 billion on services and software. That's a whopping 62 percent of IBM's gross margin that didn't come from hardware. IBM actually had higher gross margins from services and software than from hardware for fiscal years 1997 and 1996 as well.

See whether you follow my logic:

1. Based on profit motive, IBM hasn't been a hardware company for many, many years.
2. Only one thing drives shareholder value over the long term: profit. Period.
3. Corporate leaders are primarily rewarded for driving shareholder value and consequently tend to go where profits are easiest to make.
4. IBM corporate leaders have been moving IBM toward service- and software-based profits for more than 12 years.
5. Why would a software and services company want to sell something like the System i — on which everything is integrated and just works out of the box? (Oh, wait, didn't IBM recently announce the Power System, where the world of the i kind of disappears into the "p.i.—t" of generic hardware and forces us to be good at configuring and integrating things ourselves?)

We really haven't been delusional. We really have seen IBM morph into a very different company: from hardware focused to software/services focused. If this transformation began in the mid- to late 1990s, what's happening in the late 2000s? Is IBM moving ever closer to becoming a non-specialty-hardware company? A quick look at IBM's financial statements tells the story.

Figure 1 contains all kinds of hard numbers. The orange-shaded rows show IBM's revenue from the year ending 1997 through the year ending 2007. Revenue is broken down into three categories — hardware, professional services, and software. IBM separated professional services in 2007 into two categories — technical and business. For the purposes of this article, I've combined these two numbers into the single category called "Services." (Note that in each colored section of the table, the darker-colored rows simply show the services and hardware numbers added together so that they can be more easily compared with the hardware numbers.)

Now let's examine the yellow-shaded rows in Figure 1. These rows show the relative percentage of each category (i.e., hardware, services, and software) in IBM's revenue from year ending 1997 through year ending 2007.

The green-shaded rows in Figure 1 show the relative percentage of contribution of gross margin in each IBM category (i.e., hardware, services, and software) from year ending 1997 through year ending 2007. Because IBM had not yet issued its 2007 annual report at the time this article was written, I calculated the 2007 gross margin dollar amounts from IBM's January 17, 2008, financial report and then calculated the relative percentage from the gross margin dollar amounts. This is the only column in the table that did not come directly from IBM.

The blue-shaded rows in Figure 1 show the gross margin dollar amounts in each IBM category (i.e., hardware, services, and software) from year ending 1997 through year ending 2007. Again, I calculated the 2007 gross margin data based on IBM's January 17, 2008, financial report.

Now, I'm not a fan of long columns of hard numbers — I can't tell much of a story that way. So, to satisfy the visual person in me, I put together some charts based on the data. As such, Figure 2 offers a graph that shows the relative percentage of revenue by IBM operating business segments (i.e., hardware, services, software). I've purposely ignored revenue from other sources that IBM reports — such as its financing business, interest income, and so forth. As you can see, hardware as a relative percentage of total revenue (from business operations — not including IBM's financing business) has dropped from about 50 percent in 1997 to about 22 percent at the end of 2007. In the meantime, both software and services are trending up, with services trending up sharply. Services alone at the end of 2006 represented 54 percent of total sales, up about 50 percent from 1997 (i.e., 34 percent growing to 54 percent). In addition, if the trend continues, 2008 will be the first year in which hardware revenue will be in last place compared with services and software. Somewhere, Thomas Watson is turning in his grave.

The trend is even more pronounced when you look at hardware sales compared to a combination of services and software. The graph in Figure 3 shows that comparison.

If you don't see the picture clearly that IBM has long been getting out of the hardware business and is moving to the services and software business, I don't know what other evidence you might need. Remember, the red line was below the blue line (i.e., hardware was a majority of revenue) for the last time in 1997. Another observation, this one related to the 80:20 rule: Again, if the trend continues, 2008 will be the year in which hardware revenue falls below 20 percent of IBM's total revenue. Aren't we told to focus on the 80 percent opportunities and ignore the 20 percent? Hmmm . . .

Okay, so we learned the lesson that "top line" alone is not sufficient. (Remember the dot-com bubble? "All we need is market share. Who cares about profit?" That didn't work very well.) What about the bottom line? I also deconstructed IBM's annual reports to look at gross profit from its three major lines of business: hardware, services, and software. The graph in Figure 4 shows the relative contribution of gross profit, year by year, over the past 11 years.

Again, you can see that hardware's profit has dropped from about 46 percent in 1997 to a little more than 20 percent in 2007 (dropping more than half), while both services and software's contributions to the bottom line have risen substantially and consistently. What about that profit motive now? And what about the drive for IBM corporate leaders to create shareholder value by bringing in profits? We'll look later at some of IBM's forward-looking guidance, but as a preview, IBM leaders are devoted to stripping the company of its low-profit-margin products and services. (Remember the ThinkPad? IBM dropped that product line when PCs moved into the commodity space.) Another thing about capitalism is that the drive for profits is relentless. Relentless.

Okay, again to appease the "picture lover" in me, I offer the graph in Figure 5, which compares hardware's contribution to gross profit with a combined contribution coming from services and software. One more time: Where would you focus your efforts? On the place in your business where 80 percent of the profit is generated? Be honest.

But . . . the System i Has Been a Very Profitable Platform!

It is true that the System i has been very profitable. I'm sure you are starting to become a bit defensive at this point. The decline of IBM as a "machine" company is exclusive of the System i, you say. It's a great box, you might say. That's true — but is it relevant? Greatness and long-term, growing profitability are not necessarily combined — unless you happen to be Apple, and you've invented a little device called an iPod (even Apple won't be able to maintain that forever). Over the years, we've heard anecdote after anecdote about the revenue contribution of the System i to IBM's hardware business. I remember hearing sometime in the mid-1990s that the System i represented about $11 billion of revenue for IBM. I never saw a real financial statement from IBM stating that as a fact, but more than one IBMer hinted that it was approximately the right number. We've also heard things like, "The System p has never made money. Profits from the System i have been used to fund its research, marketing, and development." I don't know whether that statement is accurate, but something about it "rings true" to anyone who has been observing IBM over the years.

Now, I don't think anyone would believe that the System i contributes $11 billion of revenue to IBM's hardware business today. That said, it's brutally hard to find any concrete facts related to the decline in System i revenue growth over the years, because IBM doesn't report financial results by hardware platform. I had visions of preparing a graph showing a baseline of 1997 for the "top" (at the magic $11 billion number) and then showing a cumulative decline in System i revenue over the years. Alas, that idea foundered after several hours of frustrating searches through articles, annual reports, and other available information.

However, when examining IBM's annual reports for the past 11 years (1997-2007), I found several references to revenue declines for the AS/400, iSeries, and System i — but IBM offered few specifics. IBM did say things such as, "The 1997 decrease was primarily driven by lower revenues from System/390, AS/400 and RS/6000 servers" (source: IBM Annual Report, 1997, page 43). And "Revenue from iSeries servers declined in 2002 versus 2001 due to weak volumes across all product lines. Revenue from iSeries servers also declined in 2001 versus 2000" (source: IBM Annual Report, 2002, page 50). Further, IBM also offered the following: "This increase [in revenue] was more than offset by revenue declines for the midrange iSeries servers and the zSeries servers in 2000 as compared to 1999" (source: IBM Annual Report, 2000, page 56). In the 2004 annual report, IBM did offer some specifics: "iSeries server revenue declined 9 percent year to year" (source: IBM Annual Report, 2004, page 28). Finally, "The declines were driven by lower revenue from S/390, AS/400 and RS/6000" (source: IBM Annual Report, 1998, page 57).

IBM's 2003 annual report offered the only bright spot in reporting System i revenue: "Revenue from the iSeries increased in all four quarters of 2003 when compared to 2002" (source: IBM Annual Report, 2003, page 58). In addition, I did some Google searches on the quoted phrase "System i revenue growth" (and also substituted "AS/400," "AS 400," and "iSeries" for "System i") and got very few results. Anecdotally, there are many stories that discuss the decline of System i revenue, but I found only a couple that discussed any increase in revenue over the past 10-12 years.

Conjecture or Reasoned Argument? You Decide

At the expense of purely objective research, let's assume that the decline of the System i revenue is moving faster than the overall drop in IBM's hardware revenue for the past 11 years (overall hardware revenue has dropped about 42 percent, from $36 billion to $21 billion). If the average revenue drop is 3.8 percent per year for all of IBM's hardware revenue, I would hazard a guess that the System i revenue drop has been at least two times that. So, for the sake of simplicity, let's assume that System i revenue has dropped a total of about 77 percent over the past 11 years, or an average of 7 percent per year. That would make the business that was once worth $11 billion worth about $2.5 billion today.

If the profit contribution for the System i remained constant (let's say that the System i has always been 50 percent more profitable than IBM's total hardware group, which has shown about 30 percent profit averaged over the past 11 years — so the System i delivered about 45 percent profit instead of 30 percent profit) over the past 11 years, what was formerly a profit machine of upward of $8 billion 11 years ago (IBM's hardware gross profit was about 50 percent in 1997; if the System i of the time was 50 percent more profitable than the average then, it would have generated some $8 billion in gross margin) has dropped to $1.1 billion today. Not peanuts, certainly, but it represents a hugely different relative percentage of IBM's overall profitability. In 1997, $8 billion in profit represented about 27 percent of IBM's profits on operations. In 2007, $1.1 billion in profit represented only about 2.7 percent of IBM's total profits (an order of magnitude drop). That's significant when you ask the question "What does profit have to do with the future of the System i?"

At the time I was writing this article, IBM hadn't released its 2007 annual report or any financial statements for 2007. Therefore, the last "official" word we have about IBM's future direction is from the 2006 annual report (ibm.com/annualreport). Here's an excerpt from the "Chairman's Letter" in the 2006 report:

Dear IBM Investor: I am happy to report to you on a very strong year for our company. In my letter last year, I said that we believed we had positioned ourselves to capture the most attractive growth and profit opportunities in our industry. In 2006 we did just that, setting new records in profit, earnings per share and cash performance.

What our numbers do not reveal — and what is perhaps IBM's most notable accomplishment during this period — is that we achieved these results while fundamentally reshaping our company. Whether you look at our technology, strategy, business model, processes or culture, IBM is a very different enterprise today than it was at the beginning of the decade. We have prepared the company for growth and leadership in a radically different future — while continuing to deliver steady results. Samuel J. Palmisano, Chairman, President, and Chief Executive Officer

Reading between the lines of Mr. Palmisano's letter tells me that all of IBM's hardware platforms are at risk — especially those that cannot be successfully differentiated in the marketplace. What about mainframes? A mainframe is unique (okay, big System i machines look and act a lot like mainframes, but they aren't mainstream mainframes, and everyone knows that). People know that a mainframe came from a 370, not from a S/38. I doubt that IBM would ever think about spinning out its System z business. Doing that would rip a huge hole in the universe. By now, we know what IBM felt about a separate System p: IBM likes a generic box that runs Linux, i, and AIX — the Power System. No more System p either. System x? Blades. Period.

There will be money to be made in storage — forever — but I believe that one of the greatest profit-making ventures for IBM's Systems and Technology Group (STG — where the System i lives) will be in OEM for specialty technology. Think chips for automotive, aerospace, retail; think storage for massive needs such as weather forecasting, the human genome, and so forth. IBM will continue to shed unprofitable lines of business for the foreseeable future. If IBM was willing to cast aside the ThinkPad (with all the whining and complaining that decision created, and the potentially devastating impact to its brand — at one point ThinkPads were the laptop to own, but now? Can you say, "MacBook Air?"), what's to say IBM won't do the same thing with other platforms — such as the beloved System i?

It may not make the hard-core AS/400 types any happier, but doesn't the Power System announcement make a lot more sense now? After all, IBM can generate $1.1 billion in profit much faster by adding more consultants than by continuing to invest in a unique, difficult-to-market-yet-highly-loved specialty platform such as the System i.

Robert Tipton is managing partner of R S Tipton, Inc., and a long-time contributor to System iNEWS. His book Untangling IT: 25 Years of Lessons in Effective IT Leadership (available at the System iNetwork Bookstore at pentontech.com/education) and R S Tipton’s workshops and consulting services focus directly on the process of creating higher levels of effectiveness through innovation, inspiration, and common sense.