Gorillas in the market
Geoffrey Moore's latest book tells you how to pinpoint high-tech winners
In The Gorilla Game, high-tech marketing sage Geoffrey Moore teams up with two investment professionals to turn his theories about the dynamics of technology markets into profits for investors. Can his simplistic models and colorful metaphors translate into strategies that beat the Dow? (3,600 words)
Moore's enormously influential theories of high-tech market dynamics -- as laid out in his previous books Crossing the Chasm and Inside the Tornado -- have pushed these terms into the everyday vocabulary of technology marketing types. As a result, a phrase like, "Our company has knocked down all the pins in the bowling alley, and by next year we'll be riding the tornado on our way to gorilla-dom" is as meaningful to marketers as the phrase, "We need to tighten the code so that the footprint is small enough to reduce thrashing before we can go FCS" is to programmers.
Planet of the apes
The ultimate objective for any high-tech business, says Moore, is to become a gorilla. A gorilla dominates its market so completely that all other companies in that market -- or related markets -- are subject to its whims and strategies. Four of the acknowledged gorillas in the current market are Microsoft in desktop software, Intel in microprocessors, Cisco in network hardware, and Oracle in databases. (I'd argue about Oracle, but more on that later.)
Gorillas, however, can only exist in gorilla markets. That is, the market must have spun into hypergrowth (the tornado phase), the technology must have a proprietary architecture (be owned and controlled by the gorilla), and the customer must only be able to switch to a competitors product at a very high cost.
The classic example of a gorilla market is the so called Wintel market, based on the Microsoft Windows and Intel microprocessor gorillas. Clearly, this market has gone through the tornado growth stage and is now firmly entrenched in the mainstream. Wintel, taken as a whole, is a proprietary architecture. The Wintel market encompasses a wide range of products and services, starting with Intel's chips and Microsoft's operating systems, but it also contains computer systems by the likes of Compaq and Dell; applications by Microsoft and others; peripheral devices, development tools, file formats, consulting, training, etc. All of these are interdependent. Everyone who participates in this market is in thrall to Microsoft and Intel -- including customers, who would find it staggeringly expensive and inconvenient to switch to a competing architecture such as Apple Macintosh or Sun SPARC/Solaris.
Gorillas, however, are not the only animals in the jungle. Beneath the gorillas are chimps -- important niche players who are successful but will never overtake the gorilla -- and monkeys, which are all of the lesser competitors that end up with little or no market share. Examples of monkeys include vendors that produce low-cost imitations of a gorilla product hoping to compete on price, such as Intel-clone chipmakers and HP-clone laser-printer makers. Typical chimps include Apple in desktop computers and operating systems; Bay Networks in network hardware; and (if you believe Oracle is a gorilla) Sybase and Informix in relational databases.
In a new wrinkle for this book, Moore differentiates gorilla markets from royalty markets, in which there are clear leaders but it's relatively easy for customers to switch from one product to another. Royalty markets are often those based on open standards. Moore calls the corresponding participants in royalty markets kings, princes, and serfs -- instead of gorillas, chimps, and monkeys. The markets for most PC peripherals (disk drives, memory, monitors, modems) are obvious examples of royalty markets. It's trivial, for example, for someone with a laptop computer to remove one $100 PC Card modem and replace it with another. Therefore, no modem maker can have as much of a stranglehold on its market as, say, Microsoft does on desktop software.
From this simple hierarchy comes a simple investment strategy: invest in gorillas only. The authors claim that this investment strategy is not a high-risk approach suitable only for venture capitalists who invest in startups. On the contrary, it's supposed to be a safe way to achieve above-average gains consistently with little downside risk, making it appropriate for individual investors who are in for the long haul.
how gorilla markets behave, which
means an investor can back a gorilla
before Wall Street catches up.
This model works because, as the authors found, Wall Street does not understand how gorilla markets behave, which means an investor can back a gorilla before Wall Street catches up. Here's why: Stock prices reflect analysts' views of companies' future earnings potentials. Analysts try to predict future earnings by applying sophisticated models that have been refined over the course of decades. Moore and company purport that these models underestimate the growth that gorillas enjoy because they don't consider tornado hypergrowth. This means that gorilla stock will actually be undervalued -- even though their P/E multiples may be enormous by conventional standards. Eventually, gorillas return earnings that outdo expectations, and the stock prices adjust upwards accordingly. But in the meantime, there is a long window, typically years, in which it's possible to buy gorilla stocks and enjoy outstanding returns when the market finally catches up.
The trick, then, is choosing gorillas. This means figuring out which markets are gorilla markets, as opposed to royalty markets and markets that never "go tornado" (i.e., grow large enough to become one or the other). Next, pick which competitor will ascend to gorilla-dom. This is where things get complicated.
Moore's taxonomy of players in high-tech markets is handy and illustrative, but it only works cleanly in a small number of cases. Many factors in judging gorilla-ness are ambiguous. Take Oracle and the relational database market as an example. It could be argued that Oracle fails the "proprietary architecture" part of the gorilla test, because the Oracle database is based on an open standard, the SQL database programming language, on which its competitors are also based. As long as you write a database application in standard SQL, it's easy to port it from one relational database to another. Oracle's high switching costs are more a function of the large amounts it charges customers for software licenses, and if Sybase and Informix got desperate enough, they could respond to that by offering competitive upgrades at minimal cost to increase market share -- a tactic with a long history of success in other markets. In this scenario, the relational database market would be more like a royalty market (because of low switching costs) than a gorilla market.
In addition, vendors may have some products or product lines that are gorillas and others that are mere chimps or monkeys in another gorilla market. For example, Microsoft is not a gorilla in database server software, and its Explorer Web browser looks like a prince in a royalty market where Netscape is (still) king.
Other gorilla judgments depend on how the market is defined. Is Sun, for example, a gorilla or a king in any market? It depends. In the Unix server market, Sun is probably a king. It's the leading player, but switching costs are relatively low because it's not that difficult to port Solaris applications to HP/UX, Linux, or some other flavor of Unix. In the overall operating systems market, Sun is a chimp in the land of the Microsoft gorilla. Sun is a serf in the royalty market for desktop computers -- where the kings and princes are Wintel box makers like Compaq, Dell, and Gateway -- but a king or possibly even a gorilla in the desktop engineering workstation market. This type of ambiguity confounds anyone who would look for gorillas outside of the list of usual suspects.
Moore et al. provide three extended examples (past, present, and future) of gorilla markets and how one would invest according to their rules. The "past" example is Oracle and the relational database market, while the "present" example is Cisco and the network hardware market. Their "future" example is customer-service software, in which the players include Vantive, Scopus, and Clarify.
The problem with the customer-service software market is that it's unclear whether this market stands on its own as a tornado market with gorillas, or whether it's just part of the larger market of client/server back-office software, where the gorilla is SAP and the chimps are PeopleSoft, Baan, and a couple of others. SAP -- a gorilla if there ever was one -- is using its powerful, super-complicated architecture as a platform for customer-service offerings in certain vertical industries (such as newspaper and magazine publishing). The advantage of going with SAP for this solution is the guarantee that it will interoperate with SAP's financial applications, which are increasingly ubiquitous at large corporations. The disadvantage is that SAP implementations often collapse under their own weight of complexity and expense. The question is: Will SAP take over customer service as an application category, or will the category stand on its own? The authors raise this question, saying that it's too early to give an answer.
Another important question arises: If a category emerges and goes tornado, will it be big enough to interest the investor? Several technologies have gone into hyperdrive but without making money for shareholders -- or without having any public shareholders at all. I'm not talking about technologies that dominate small niche markets, either. Consider, for example, the many technologies which achieved ubiquity by being bundled into Microsoft DOS and Windows, like disk storage compression, full text search, and faxing software. Many of these started out with the potential to spawn gorilla markets, but they couldn't come up with a way of charging customers enough money to be big, or they simply got sucked into the Microsoft juggernaut. Moore et al. don't address this type of vendor at all.
Searching for a gold mine
In other cases, technologies that become universal take unexpected steps during their growth. If there are a handful of competitors in the early stages of the market, it's not necessarily a given that they will emerge as money-printing gorillas and chimps. Sometimes other players come along and figure the opportunities out correctly.
One example of this is full-text search-engine technology. A few years ago, I thought that text search would be big because it seemed the obvious answer to access the fast-growing amounts of information on hard disks and large public and corporate networks. (Web technology turned out to be another answer, but leave that aside for the moment.) So I invested in two companies that were public, had leading text search engines, and derived most or all of their revenue from them: Verity and Fulcrum. In retrospect, text search had at least some of the trappings of a gorilla market in 1994-95.
The only remaining condition of
gorilla-dom was that the market
needed to spin into hypergrowth.
It did, but none of the early vendors
made much money from it.
For one thing, the technology resisted an attempt to be based on open standards. There was an open standard protocol for distributed text search with the catchy name of ANSI Z39.50, but it achieved spectacularly little acceptance among text search vendors. In fact, the only vendor that embraced it was its inventor, Brewster Kahle, and his company, WAIS Inc. (which was eventually acquired by AOL). Instead, each of the roughly half-dozen leading text-search vendors developed proprietary architectures for distributed text search. Switching costs among these competing technologies would be high because each vendor's data format was also proprietary, and executing data conversions on millions of pages of digitized text was not a simple matter, either.
The only remaining condition for gorilla-dom was that the market needed to spin up into hypergrowth. It did -- text search is now absolutely everywhere. But none of those vendors made much money from it. Verity's search engine got bundled into Lotus Notes and a few other products. Fulcrum licensed its product to Microsoft. WAIS developed a "lite" version called FreeWAIS that became popular in Web sites but, as its name suggests, was given away. Other vendors like PLS and Excalibur made modest businesses in high value-added niche markets.
It took the Web to turn text search into a generator of profit through stock shares. Text search engines now support Web sites that make money from advertising revenue, such as Yahoo and all of the "Yah-toos."
At least two of the original text search engine vendors tried making transitions to the Web. One was Verity, which stayed with its model of selling tools, this time to Web site builders. Another was a startup company called Architext, which had a late offering in the search engine sweepstakes with little to differentiate it from the existing players. Before going public, Architext changed its strategy ... and its name. It is now called Excite -- and, as Paul Harvey says, now you know the rest of the story. Incidentally, Moore et al. characterize the Web portal market as royalty, not gorilla/chimp/monkey, because of the extremely low customer switching costs. And by the way, I eventually sold my Verity and Fulcrum shares at a loss.
Reading between the lines
Moore and his co-authors don't address the relationship between growth and profitability. They spend a lot of time toward the end of the book explaining how to detect gorillas. They list many information sources, starting with the basic periodicals for individual investors who can only spend a few hours a month doing research, and ending up with the expensive "boutique" journals for professional investors. They rely heavily on the media's perception of technologies and vendors to determine gorilla status. They argue successfully for a "herd" mentality, i.e. that the media perception of technology tends to feed on itself, and hype, therefore, becomes self-fulfilling prophecy.
Trade magazine headlines, for example, are an indicator of possible gorilla-dom. "Fueling the Blivitz Juggernaut" probably means that Blivitz may be a gorilla or royalty market in the making; "Pretenders to the Blavitz Throne" means that Blavitz is probably a king or gorilla; "Blevitz Reviving its Dusty Image" means that Wall Street has finally caught up with Blevitz, and it's time to sell.
In addition to tools for identifying gorillas, The Gorilla Game gives a set of fairly precise rules for investing in gorillas. The rules differ according to whether the technology is a software application or an enabling technology such as hardware or operating systems. The rules are simple, clear, and straightforward -- and don't require daily exercise. On the contrary, they call for you to make buy-and-sell decisions on a quarterly basis.
you think it's going to be a gorilla
market, you buy because it is a gorilla
market. You give up the astronomical profits
that are possible if you guess right before
the tornado hits, but you also eliminate
most of the risk of guessing wrong.
The point, once again, is that theirs is a conservative investment philosophy. You don't buy into a technology because you think it will be a gorilla market (as I did with text search engines), you buy because it is a gorilla market. You give up the astronomical profits that are possible if you guess right before the tornado hits, but you also eliminate most of the risk of guessing wrong. The authors apply their rules to the three scenarios as conservative, nonprofessional, individual investors would, and the results show that you can still make serious profits this way -- because it still takes Wall Street longer to recognize gorilla growth.
The Gorilla Game is a good book, if not as fun to read as Moore's previous works. (Moore really is a gifted writer; his co-authors aren't.) It's likely that the investment strategy it espouses is a good one, but as I read it, I continually came back to the same question: Will gorillas and gorilla markets continue to behave as they are now vis-á-vis Wall Street's valuation of them?
In other words, the gorilla investment philosophy depends on how gorilla markets develop and how Wall Street reacts to them. What if this changes? For example, as more and more marketing people understand how gorilla markets work, they will spend more and more effort to undermine them.
What if this knowledge increases to the point at which gorillas are uncommon and/or only remain gorillas for short periods of time? At that point, the dynamics of markets will change, making the assumptions on which the gorilla investment model is based invalid. As another example, what if this book (and others like it) is so influential that it changes the way stock analysts value technology companies so much that stock prices more accurately predict gorillas' future earnings? For example, it seems fairly obvious that someone at a mutual fund company, upon reading this book, will become inspired to create a fund based on its principles: The Gorilla Fund. Paul Johnson ought to be able to make this happen at his employer, BancAmerica Robertson Stephens.
I myself am scared of the stock market right now. After having cashed out over half of my holdings to buy an apartment, I quiveringly hold onto a small number of blue-chip growth stocks as the market gyrates. Nevertheless, I would probably be tempted to invest in The Gorilla Fund, as long as the authors of this book are actively involved in picking stocks. Their rules for picking gorillas may be somewhat hazy, but I think I would trust them to do the picking for me. That's the most important value these authors could offer.
Title: The Gorilla Game: An Investor's Guide to Picking Winners in High Technology
Authors: Geoffrey A. Moore, Paul Johnson, and Tom Kippola
List price: $26.00
About the author
Bill Rosenblatt is market development manager for media and publishing industries at Sun Microsystems Inc. Reach Bill at firstname.lastname@example.org.
In my review of Quittner and Slatalla's Speeding the Net in July's Bookshelf column, I referred to the idea of freeware, as espoused by Richard Stallman of the Free Software Foundation. Freeware is based on the notion that intellectual property should be a public good, as opposed to a source of private gain. I characterized this idea as "Marx-derived," which the editors of this magazine changed to "Marxist."
Stallman himself has contacted me to deny that Marx was in any way a source of his inspiration, calling my characterization "harsh." He says, "I don't think Marx can claim credit for the idea that people should help their neighbors; that idea is many centuries older than Marx."
I had thought I was familiar enough with Stallman's ideas, having contributed to his highly successful GNU Emacs editing system and having written a book about it. But apparently I'm not. Although I did not intend the term "Marx-derived" (or even "Marxist") to be derisive or negative in any other way, I regret the mischaracterization and welcome the opportunity to retract it. Stallman's ideas have been inspirational to a generation of software developers, no matter what their intellectual antecedents may be.
If you have technical problems with this magazine, contact email@example.com