Sunday, January 20, 2008

The Acquisition Candy Store, Part 2

Oracle's three-year series of acquisitions, including most recently BEA Systems, appears to be the implementation of a "downstream" vertical integration strategy.

The underlying rationale (or "investment hypothesis") for this strategy might follow a chain of thought such as:

  1. Oracle's core business, historically, has been the development and sales of relational database systems.

  2. Oracle has attained an outstanding competitive position in the RDBMS market by excelling at its software development and sales processes.

  3. Although the RDBMS market continues to grow, the market's growth has been slowing. Consequently there is less need and less opportunity to reinvest in the core business and instead Oracle now finds itself with a surplus of cash.

  4. More alarming, Oracle's ability to grow profitably in the core RDBMS market is at risk. Open source and lower priced competitive products threaten to commoditize the database technology, eroding Oracle's pricing power in its core business.

  5. There are no obvious ways to reinvigorate the core business, and so Oracle's best opportunity to continue growing is to find a "business adjacency" with greater profit potential than the core.

  6. Oracle sees "downstream" vertical integration as the most logical business adjacency to pursue. Specifically, the enterprise applications software that runs on Oracle's databases, and the middleware that interconnects Oracle's databases with other corporate systems.

  7. Oracle believes it is better to dominate the enterprise and middleware segments than to just participate selectively. (Perhaps Oracle believes there is value in being a single-point-of-contact vendor, or Oracle wants to create larger closed systems, or maybe Larry Ellison just wants to win here as badly as he wanted to win in RDBMS).

  8. Acquisitions are a much more effective way to enter this segment than internal software development. A number of high-quality companies had already staked out positions in enterprise software and middleware, and it would be much faster and easier to acquire them than for Oracle to attempt to develop alternative products themselves.

  9. Oracle therefore has launched a campaign to identify and acquire best-in-class software companies in all aspects of enterprise software and middleware.

Viewed in this light, Oracle's acquisition spree seems to make very good sense.

But it raises some questions as well.

To what extent is Oracle integrating its acquired companies into a common corporate structure and culture, or to what extent is Oracle continuing to let each one exist on its own? What value is Oracle adding to these companies by being their parent? Will customers prefer the one-stop-shopping approach that Oracle essentially will be offering, as opposed to selecting products from multiple vendors? How much do these moves threaten Oracle's other business partners, against whom Oracle is now competing more directly? Will Oracle be able to maintain the pace of technological innovation in its acquired companies, or will these businesses become less competitive?

All topics for a later time.

Thursday, January 10, 2008

The Acquisition Candy Store, Part 1

I once had a conversation with a software company executive about capital allocation. Like many successful companies, his company was generating cash hand-over-fist. As part of his job, he recommended major capital allocation decisions.

We talked about the various options they had, and a little bit on how he thought about the returns on investment for internal projects versus business acquisitions.

Then I naively asked, "What about when you don't have any good reinvestment opportunities? Do you ever look at the return on investment of share repurchases, or even on paying a dividend and letting your shareholders reinvest somewhere else?" (You see, back in the Old Economy, where I work, we actually ask ourselves questions like this. Our shareholders ask us this too. A lot.)

Well, he gave me a look that blended the most profound, heartfelt feelings of sorrow, pity and regret. As if his hope for having an intelligent conversation with me had been cruelly and abruptly extinguished. His smile and enthusiasm collapsed. You poor, misguided fool, his features seemed to say. Did you really believe that crap they taught you in business school?

Needless to say, I found a way to change the subject. But it left me there something about big, successful software companies that compels them to want to keep all the cash they generate, rather than give it back to their shareholders?

Now the executive I talked with most definitely did not work for Oracle, but Oracle might be a good illustration of this particular behavior. ORCL is certainly a reliable cash generator:

And, starting in 2004-2005, Oracle went on a tremendous spending spree. Not just the big name acquisitions, but a number of smaller companies as well. Look what happened to their goodwill line:

What was Oracle's software strategy here? Were these dozens of acquisitions made to support their business strategy, or were they seen as the most sensible way to keep the cash in the family?

Saturday, January 5, 2008

Will TopCoder shift the balance of power from employers to programmers?

One of the things you learn behind the scenes at business school is that there are two ways to make a lot of money (not counting winning the lottery or being born into the right family!). The first way is the "business leader" can start a company or be a top executive, preferably a CEO. The second way is the "talented expert" can excel at a measurable skill that someone is willing to pay a lot of money for. For MBA's, this used to mean investment banking. These days it more often means hedge funds or private equity.

The "talented expert" route to wealth appears in many different fields. The top NFL players make more than the owners. The top actors make more than the studio executives. The top hedge fund managers make more than the CEOs of investment banks. (IIRC, this phenomenon was documented in The Winner Take All Society).

Aside from several highly publicized Google poachings from Microsoft, we haven't seen as much of this happening in the software industry. Most of the top talent in software development is hidden. It's difficult for any one talented programmer to prove to the market or to an employer that he or she is worth ten times the compensation of another, less talented programmer.

And that's how it has always been, until TopCoder came along.

TopCoder is transforming the world of software development in more ways than one. Not only is it assembling a massive virtual network of programmers from around the world, it is also providing objective, quantifiable ratings of their skill levels and demonstrable evidence of how they work together in teams.

Think about this. The balance of power may be shifting more toward the talent and away from the employers.

Thank you to Judith Hurwitz's Weblog for prompting this thought. And thank you to TopCoder as well, where I have been an occasional competitor since 2005 using the handle JRR.

Tuesday, January 1, 2008

RDBMS Market Share

Three vendors, Oracle, IBM and Microsoft, control over 80% of the relational database market today, and the same three have led the industry for at least the past decade.

The above chart is based on press releases from or quoting Gartner Group/Dataquest. For specific references see the backup spreadsheet. Note that up until 2004, Gartner/Dataquest only included new license fees in its market data, then began including subscription and renewal fees in 2005. So the data for 2005-2006 is not strictly comparable with the data for previous years.

The shift in mix from 2004 to 2005 because of the reporting change also implies a difference in the major vendors' business models. The fact that Oracle "gains" share and IBM "loses" share when renewal and subscription revenues are added in implies that renewals and subscriptions are a more important part of Oracle's business than of IBM's. (Or, perhaps, that Oracle is consistently willing to lowball IBM on up-front fees in order to make the sale!)

RDBMS Market Size and Growth

The relational database market grew from about $10 billion in 1998 to about $16 billion in 2006, a compound annual growth rate of 7%.

The market research firms IDC and Gartner Group publish annual reports on the size of the market and the shares of the major participants. The latest IDC report costs $4,500, and the Gartner reports are also far from inexpensive, but the two firms do provide some free highlights in their press releases. Oracle has also republished certain IDC statistics. Using this information, it is possible to piece together a picture of the market's size and growth:

These figures include revenues from new installations, as well as renewal and subscription revenue for existing installations. In prior years, Gartner (under the name Dataquest, a market research firm that Gartner acquired in 1995) would typically emphasize only the new installation revenue, whereas IDC would typically emphasize the combined revenues.

This chart also suggests that the tech crash had a relatively modest effect on the database vendors' revenues. Yes, they all had to cope with a flat to slightly shrinking market rather than a growing market, but as industry contractions go, this one appears fairly mild peak-to-trough compared with, say, network equipment or basic materials. The market plateaued during 2001-2003 (the dip in 2003 may be an artifact of how this chart was assembled from IDC and Gartner data; the actual bottom more likely occurred in 2002), but by 2004 the market had reached a new record high.