Information technology [IT] marketeers thrive on inventing buzzwords. But IT researchers know that it is hard to get reliable IT-consumer data about a “whizbang whatsis” when the term was just invented by an IT marketeer shortly before the interview. It goes like this:
Q. Which vendor’s “whizbang whatsis” do you use?
Q. Do you have a “whizbang whatsis?”
A. What is it?
Q. Do you plan on installing a “whizbang whatsis” next year?
......And so forth.
Statistical data derived from market research instruments that require you to explain your question is of very little value to the developer, marketeer, or investor, except typically to prove that no one knows what a “whizbang whatsis” is. [Truth in advertising: I enthusiastically participated in foisting heuristic operating system, integrated office automation, and application service provider [ASP] on the IT market in the ‘70s, ‘80s, and ‘90s respectively.]
So, in now doing IT investment research, every time I hear of a new “whizbang whatsis” I ask myself:
• Is there incremental revenue opportunity in the IT market for this new “whizbang whatsis?” (as there was for spreadsheet processors, for example, although at the expense of green ledger paper and #2 pencils)
• Or is this “whizbang whatsis” just a new marketing term for some IT for which there is already a market and therefore a zero-sum revenue scenario will be in effect? (as was the case for client/server-design packaged applications, which slowly replaced monolithic-design packaged applications over a 20-year period)
• Is there a particular company that can take advantage of the “whizbang whatsis” marketing buzz even if there is no incremental demand? Assuming the buzz takes off! (as was the case with PeopleSoft in client/server-design packaged apps)
• And if so—if there is no incremental demand and a zero-sum scenario is in effect—what companies are going to be hurt as other companies benefit? (as in PeopleSoft displacing McCormick & Dodge, Cyborg, and so forth)
“Software appliance” is the “whizbang whatsis” of 2007.
• Business Objects (BOBJ) didn’t announce a software appliance last week, but still got some buzz because it announced the inevitable “Open Appliance Alliance.” (They didn’t really call it that but I can’t help myself.)
• Ingres announced a software appliance last month.
• In December IDC predicted “software appliance” would become a household word in 2007.
• Ex-Red-Hat exec Billy Marshall blogs on the subject frequently. [Further truth in advertising: Marshall is now CEO of rPath, which provides a “software-appliance platform,” and which belongs to Business Object’s appliance alliance.]
So what is a software appliance and what are the answers to my IT investment research questions about this particular “whizbang whatsis?”
Well Wikipedia says (I paraphrase) it’s an application pre-integrated with infrastructure software such as web server software, a file system, and operating system in one easily installable and manageable stack. I can understand that. It’s back to the future to the GE Model 58 40 years ago, any number of minicomputer OEM mid-market enterprise application offerings (particularly on the AS/400) in the 1980s, and an even shorter journey back to the ill-fated SAP/Intel-sponsored Pandesic B2C ecommerce package in 1996/1997 (the first version, not the later morphed-into-an-ASP version of Pandesic). I don’t mean to put a curse on software appliances with that last analogy.
The idea of pre-configuration and self-maintenance has always made sense where enterprises—particularly mid-market companies—need lights-out, look-ma-no-hands operation. Unfortunately, if I am reading Wikipedia correctly, the Wikipedia definition’s author is also a Billy Marshall so I needed to get a second opinion.
Two years ago before Marshall’s rPath existed, Business Signatures [BS] said it was pioneering the idea of software appliances. I find no earlier reference to the buzzword on Google other than in the context of the more familiar hardware-based “security” or “storage” appliance so I’ll give BS the nod for first mover. BS proposed five characteristics in this NetworkWorld article at that time. The BS characteristics are a superset of Wikipedia/Marshall’s definition. BS adds a particular type of pricing and a standards filter to the definition. Using descriptors such as pricing models and the need to be standards-based is also not good in IT market research:
• They have nothing to do with functionality and design.
• They do not determine market success or failure.
• They are usually a means one or more vendors use to say “no one else can play.”
Interestingly, although almost no one was talking about software appliances before BS in mid 2005, by the time Ingres said it was working on one in early 2006 (the product that was announced a few weeks ago), InfoWorld found no need to explain what a software appliance was. That’s how quickly a buzzword can take hold.
But like the buzzword ASP in the late 1990s, in my opinion “software appliance” is going to go to buzzword limbo to join “artificial intelligence,” “eBusiness,” and “web services architecture.” And sort of for the same reason. Users quickly said, “Isn’t an ASP just a service bureau?” And now users are saying: “Isn’t a software appliance just a package?”
To me, the software appliance buzz is all about packaging. Pre-configured and/or embedded software has been around almost as long as the packaged software market. The GE Model 58 accounting system that I worked on in 1970 came with everything but blank punched cards. J. D. Edwards’ first System/38-based packages, the Genuine Auto Parts DG/NOVA-based point of sales systems that went into thousands of NAPA stores in the 70s and 80s, early Great Plains Software Btrieve-based products, and other software clearly met the Wikipedia software appliance definition. They were “installed,” “administered,” and “managed” as a unit by the accountant, retail clerk, or CPA that used them. Until the PC era of course, such "appliances" even came with the hardware (but back then that was considered a limitation, not a strength).
That doesn’t mean the idea of investing in a “software appliance” supplier is flawed. But buzzword weakness is a strong indicator that the IT market opportunity is not incremental. Although a zero-sum scenario is in play with software appliances in my opinion, some competitors might take advantage and cut into other suppliers’ market share. But I just don’t sense any traction for anyone.
• Business Objects’ effort actually mates long-time query-tool market leader Crystal Reports with the ultimate in hardware appliance packaging, the Netezza Data Warehouse server. So Business Objects is unlikely to take share from anyone since it is already so strong.
• Can Ingres grab a percent of share from Oracle (NASDAQ:ORCL) with its software appliance idea? If Ingres gets more than one Icebreaker deal in competition with Oracle, Redwood Shores will be all over the software-appliance buzzword the same way it jumped on SQL 30 years ago. (One of the problems with non-incremental buzzword markets is that the cost of entry to competitors is very low; a few press releases and “you too can be marketing a software appliance.”)
• Another software appliance is replacing PBXs and another electronic medical record management.
• SugarCRM is available LAMP-stacked in addition to on demand, on site, and as a hardware appliance. Having four package choices is good but hardly likely to make someone select SugarCRM over Siebel.
• Ex-Transarc and BEA technical guru Scott Dietzen is playing in the space, making software appliance an optional packaging mode for his new collaboration software, called Zimbra. Scott’s acceptance of an idea is always a plus in my opinion and his Zimbra white paper (.pdf) adds another characteristic to the Wikipedia and BS definitions, which says that with software appliances, users only have to deal with one vendor (a good feature but also not researchable).
As for the rPath product idea that is marketed as a software appliance platform, as opposed to an appliance itself, here’s what I see. It’s marketed to handle check-in/check-out, configuration, back-up, provisioning and license management of an ISV’s appliance as well as support for Linux and virtualization. Except for the last two features, that sounds a lot like a 1990s CASE tool, just to mention another term from the IT buzzword trash pile. rPath says that its products are more about “binding” the IT pieces together (especially in virtualization scenarios) than about computer-assisted systems engineering. Whichever, because CASE tools did not take off as an IT category, this may be an incremental revenue opportunity. But just was with Oracle vis a vis Ingres in appliances, tools suppliers such as Borland and others will quickly be all over any platform opportunity if it takes off.
Disclosure: Author has no position in any company mentioned