Skip to main content

Nobody Thinks They're An Enterprise

Posted by timboudreau on August 18, 2008 at 12:56 AM PDT

I wrote this blog nearly two years ago, and was politely asked/advised not to publish it. If I rewrote it today, I'd probably make it a bit shorter. But I think the points are ones our industry needs to learn. The important one is very, very simple: playing to the ego of a non-person is a really lousy marketing strategy.

So, today's the day I publish this rather benign rant, and either ruin my career, or more likely, wonder, "what was the big deal about this anyway?". I promise I'll get back to talking about software architecture soon :-)

2006-09-06 22:08:05

The first question I was asked in today's NetBeans Day got me thinking about an unsaid thing that needs to be said. I've been thinking it for years. The entire technology industry's marketing organizations are focused around something that is more fantasy than reality. Hoisting myself onto my soapbox...

The question was: "Can I use the Enterprise Pack for NetBeans if I'm not an enterprise?" The answer is, of course you can! The fact that the question had to be asked is a symptom of a disease endemic to the technology industry. It can be summed up in one word: Enterprise.

My first close-up encounter with the enterprise cargo-cult was in 1999, when NetBeans was first acquired by Sun. Some folks came out to Prague, and kept talking with fear and awe about enterprise customers (best spoken in dramatic basso profondo, ala James Earl Jones or similar). The impression it made was that an enterprise customer is most like an irate two-year-old with teething pains, a baseball bat and the muscles to use it. I'd never heard of these terrifying enterprise customers before - I'd generally just been dealing with people in our beta program who worked for big companies and gave us great feedback. But apparently the greatest horror that could befall you was to talk to your customers - because then one day Scott McNealy might be out playing golf with the president of Smrdistan, and he'd say "I need you to fly that programmer, Jan Pi?ovak, out to fix my technology problem!" and then you'd lose one of your best programmers for a month. This was such a horror that the best way to avoid it was to make sure programmers never actually talked to the people who use what they make. It would have all been funny if it weren't so very, very weird.

I thought to myself, how did these people stay in business? They think our customers aren't people.

Enterprise is a market segment. The thing about market segments is that you (as someone marketing something) segment the market. That is, you're the one with the knife carving up the world into segments and naming those segments. The problem happens when you start thinking the rest of the world knows your names for them, or that you are talking to a segment, not to actual people. And the height of silliness is to think that the names you made up for market segments will actually mean something to the people in those segments.

What I don't get is how this point is nearly completely lost on the technology sector. Google sometime for enterprise software. Contemplate the amount of money spent on all the web pages and sponsored ads you see. Have you ever been looking for a piece of software or technology and that was what you searched for? Ever wake up one morning and say to yourself "Gee, I think I'll get some enterprise software today! Dunno what it's for, but I'm an enterprise, so it's the software for me!"? I bet you looked for something related to an actual problem you wanted to solve.

But what is a poor marketing drone to call these potential customers (bear in mind I've been a marketing drone, but hopefully not too drone-like and never for something I didn't believe in)?

Enterprise works. It has heft. It's one word. It sums things up nicely. But it's a segment. It's your name for a bunch of people. Toyota doesn't make a car called a Toyota Soccer Mom. What an unbelievable opportunity they're missing...not.

One of the best essays or statements I've ever read is Zed Shaw's Indirection is not Abstraction. That's dead on. The (basso profundo) enterprise is a layer of indirection - it's a thing that makes it harder to talk to people. The confusion is mixing that up with an enterprise being an abstraction for (and thus an okay substitute for) people.

Now, in any other industry, you'd go out and figure out how to reach those people and if you need to adjust what you make or how you make it or how and where you communicate to reach them better, you'd do that. In technology, you go up on the rooftop and shout EnterpriseFooFaw and wonder why all the cars in the street didn't just screech to a halt. And if they don't screech to a halt, you go out of business. Or go find a job at the next bit of fresh technology meat that comes along and run it into the ground too.

Soccer Mom is a useful market segment. It's also a near-insult. There's a reason we will never see the Toyota SoccerMom or the Ford SoccerMom driving down the street.

Java Studio Creator was a good tool. But it was specifically targeted at something called corporate developers (code for people who program in Visual Basic), and so was the marketing campaign. Way to lose your audience. I once asked a room of about 30 tech writers to raise their hands if they were a corporate writer. Big surprise, nobody did. Could somebody find commonalities among those people such that they'd probably really have a use for, or imagine they had a use for, specific products? Sure! Would insulting them to their faces help sales? Probably not (I got in big trouble in that preso for a slide with corporate developer surrounded by a big red circle with a slash through it - there were people who actually thought telling people they were corporate developers was a Good Idea™. While we're at it, why not substitute drone for developer? - but that brings us to a whole other topic, which is how corporate developer plays well to "analysts&quot [i.e. the people who determine your stock price] even if it makes no sense for your customers - which is its own industry-wide problem - talking to corporate analysts might be one step closer to talking to your customers, but there are a few thousand steps that come next).


I've seen quite a few successes and failures in technology marketing over the years. The successes tend to be targetted at people and specific problems; somewhere deep in the Matrix-like enterprise fantasy pod farm is a human being who will decide to use or not use what you make. We're not talking to the Borg Collective here. The spectacular failures tended to be cases where (aside from genuinely lousy technology - there are plenty of ways to blow it) they were promoted to the vendors' fantasy of their customers, rather than to the actual people who actually are trying to solve a problem the technology would solve, and who were capable of making a decision.

We can't control stupid marketing campaigns, but we can constrain them. So here's Tim's Rule #1: Any technology whose name contains the words business or enterprise is crap until proven not. And might be crap even when proven not.

If enough people take that to heart, maybe we'll see a few less really stupid marketing campaigns.

People make technology; people use technology. It's really people all the way down. And that's the "message" of this blog: Nobody thinks they're an enterprise.

2008: It's probably clear here that a lot of my criticisms and comments above have a lot to do with Sun. So I should give credit - since I wrote this, Sun has open-sourced the JDK and quite a few other things - there is a real turn-around happening in terms of connecting with people who use what we make. But there's still a long way to go.

Related Topics >>


Nice post, Tim. I can't figure out why anyone wouldn't want you to post it. I always felt that "enterprise" was a code word for "big enough company that they'll buy your product even if it's crap, and are willing to pay far too much for it". That usually coincides with having an IT department that's big enough that individual people in the company don't have a say. While at a large company, I'd often use some software that I thought was worth a hundred bucks or so, only to find that the IT dept was paying $10000 per seat per year for it. And I'd find out that we were paying huge sums for support for something like Solaris or Java, and no one in the company (except the IT dept) even knew we had a support contract. Toyota won't name a car the SoccerMom, but Lexus certainly might name a car the Elite, or the new Cadillac UpperCrust. Look at the names of the frequent renter programs ("OnePass Elite") or credit cards ("gold"). Branding with "enterprise" is trying to make it seem sturdy/scaleable (e.g. Red Hat Enterprise Linux).

Nice post, Tim. Hope it didn't get you in trouble, but I doubt it did. So you probably don't know about my history @ Sun in the early days. I was the very first guy who tried to move Sun toward doing software for servers. An unheard of, laughable thing in 1988. But that's what they needed to do, and eventually that's where they went. Took me 7 years to get the first generation of not-very-good software for the (cringe) commercial customers out there. Everything started from the customer up. The word "enterprise" was never uttered by me. Seriously, not in 7 years. But that's because I knew what I was doing. Once we actually tried to sell the stuff, the only way to get through to the sales force and the industry analysts was to use the industry-standard shibboleths. People who didn't know what they were doing just piled on, and the Enterprise wording started creeping in. Mainly, it was Zander, but there were plenty of PR wonks who amplified the noise. Your post is right, but it's also wrong. Big companies can't think. Only individuals can (the ironic inward twist of your thesis). Big companies simply react to outside stimuli, even if their own actions caused the original stimuli (this is Dilbert-esque to watch). And Those Who Matter in the industry (wall street in particular) tell you what you've got to say. "Enterprise Software" is essentially the creation of IBM and Oracle and SAP, designed to make sure that smaller companies are kept out (particularly Microsoft, but a thousand other long-dead competitors). The Enterprise Customer (let's say Fortune 500) actually does have an amazing array of common characteristics. They aren't a bad segment to go after from a corporate perspective. Unfortunately, companies try to skip the hard work of figuring out what the customer requirements really are (because this is expensive to do), and the PR wonks make up vague generalities because they don't have any real facts. (Even if they had the facts, they'd be too shy to actually proclaim them.) Of course, everyone should go re-read The Cluetrain Manifesto. You might also want to check out

This post submitted by a human being, not an Enterprise.

What I'm wondering: In how much would a government agency be/not-be an enterprise? I know, some people might simply salivate at the thought of a government contract. That's not what I'm asking about :^) PS: Great article. I really appreciate this level of candor in regards to business -- and yet more so, frankly, when the candor is accompanied with substantial knowledge, like as in above. I'd like to think the real P.T. Barnum would be proud ;^)

Wicket sounds like it's worth checking out. Granted, JSF is simpler than oracle's ADF/UIX but only if you avoid the tooling aspect. Both suffers from the fact that they were designed to make it easy to develop a new system. However, the world breaks down when you want to maintain and refactor because you need monstrous tools to make sense of it all - and tools rarely handle versioning very well. I'm more and more convinced that tooling should be an assistance, not a necessity. Which is what scares me a little about SUN's newly found focus in this aspect.

This is one of the most cogent things I have ever read. Now please excuse me, I have to get me some of that Enterprise Software.

I know what you mean, having spent the weekend learning more than I ever wanted to know about portal servers by diving into the source code to try to figure out why some very simple and straightforward things don't work. I get more and more impressed with the simplicity of Wicket. It's kind of a pity. Our industry is designed around making things complicated so that it requires an army of consultants to do very simple things (I think some of this comes from irrational fear of having to actually read someone else's code and being willing to do so is a quick way to get a leg up in this industry - and you can do it quickly - just figure out what must be under the hood for the software to do what it does, then test those assumptions and grow your perspective when you find someone actually found a different way to do something - reading other people's code is vastly more valuable than writing code.

My hat off to speaking frank when nobody cares to do so, especially considering who your employer is. I think it's worth considering why we have these marketing schemes. In reality, the monikers of "simplification" and "consolidation" really looks like this: Solution -> complexity -> problems -> work -> money -> power! Do we really need more proof of this than contrasting what Google does vs. what Oracle does? Now excuse me, I must divert some attention to the EJB3 "people beans" I have cooking.