Why is Borland’s SDO Marketing So Misguided?

Back in August of 2004, a lot of people took note of this blog entry from Danny Thorpe promising, "Disruptive ideas ahead" to be revealed at BorCon. And indeed, I thought that the BorCon presentation did not disappoint. Boz Elloy noted that the majority (!) of software projects fail and compared the business of software development to being "the cobbler’s children," lacking the process automation tools we software developers have delivered to other industries. Promising better solutions, Boz and other Borlanders followed this up with a great skit illustrating how developers, development managers, and others might benefit from the first phase of a new approach. He called this approach "Software Delivery Optimization," or SDO.

Just so I have your attention: The ideas are explosive, the products are in the pipeline, and any developer who actually wants to enjoy themselves and pay their bills writing software can’t afford to miss them.

One critical thing to note at this point is that business realities are not somehow the exclusive concern of large companies. It doesn’t matter if you are a single developer, consultant, or an employee of a small company or a mega-corporation. If you write software for a living you need to be aware of the business as well as the technical aspects of what you do. At my company (a small ISV) we practice open book management, where every employee has the right to see full company financial data and the expectation that they’ll participate in keeping the business on the right track. And if you work for a smaller company — say, a sole proprietorship — then you must deal with the business of your products, or you won’t have a company anymore.

Shortly after BorCon 2004, Borland released Delphi 2005. While not an SDO product per se, it has a lot to offer from a process-improvement point of view. Most importantly, Delphi 2005 Enterprise and Architect include StarTeam licenses, and support for CaliberRM. Since Borland purchased StarBase a few years back, the cost of entry for a basic StarTeam system has plummeted from over five figures to being free with the purchase of Delphi Enterprise! Then Borland released CaliberRM 2005, which for the first time included EstimatePro as part of the product. While all of this arguably doesn’t press the boundries of standard Application Lifecycle Managment [PDF] too far, it does give a sense of where Borland is heading at the beginning of the SDO push: Better integration between tools, making a subset of features from formerly expensive products available at lower prices, and combining products to meet the needs of specific people/roles in the business of software development.

Remember, though: These tools are just the first baby steps towards phase one (of three) of the SDO plan. If that was all Borland intended to deliver than Danny’s rhetoric would be overstated to say the least.

But it isn’t. The truly innovative stuff is still in the pipeline. So what is it? And how will it help you as a developer, an architect, an analyst, or a tester? Who else should be interested in this? Those are interesting questions, and I don’t thing Borland is doing a good job of answering them. Actually, I think they’re doing a terrible job of answering them.

Maybe you’ve visited Borland’s SDO page. Maybe you’ve visited Borland’s Core SDP page. Maybe you’ve wondered why they don’t link to each other. I find it all a bit baffling, personally, and as someone who talks to Borland executives and R&D team members from time to time, I suspect I’m better-informed about what this stuff really is than the public in general. Can anyone look at that page and state with certainty whether it is describing new products, new features in existing products, a process, or all of the above? When I click the architect and developer roles (since they’re the most relevant to what I do), I see a lot of stuff about Java. Does that mean that "Core SDP" is only for Java? What does that say about SDO?

Just when I thought it couldn’t get any worse came the recent changes to the Borland Developer Conference tracks. One would expect that SDO would be prominently featured, as it seems to be the principal focus of Borland Software Corporation right now. Sorry:

Although, some coverage of SDO and enterprise products will be included in the Conference program, a separate Enterprise Essentials Conference is also planned with more complete coverage of these topics.

I’ve criticized the handling of this year’s conference in the past, but one of the things that many people like about it is the strong developer focus. So the message above is baffling. Even though I know it isn’t true, I read the above as saying, "Sorry, guys; our biggest priority right now just isn’t relevant to you. Come to the Enterprise conference if you want to swim with the big fish."

Update (23 March 2005): To be clear, I’m not opposed to splitting the conference. I like the idea of a developer conference, too. I’m only suggesting that SDO should not be excluded from a developer conference.

Most frustrating of all for me personally is that I know that Borland can do this better. I’ve seen it done, on stage at BorCon. The message to developers needs to be loud and clear: SDO will help you stop wasting your time on ill-conceived cruft the market doesn’t want and make money doing what you love.

Sounds nice, doesn’t it? Of course, it needs to be said in a way that people don’t write it off as marketing fluff, and I have some (hopefully) constructive suggestions for that:

• The video of the BorCon presentation could be made available for viewing on BDN. People liked that. Update (26 April 2005) Borland has now done this.
• The fact that the BorCon presentation was well-received demonstrates something important: Developers want use cases, even if they don’t know that they want them. That’s hard to get from white papers, which seem to take an approach which is more theoretical. For example, when I read this document [PDF] I don’t see the kind of example of how this makes the developer more able to respond to real-life issues that I got from the BorCon skit.
• Developers tend to think in terms of products, even when they should be thinking about process. This observation is important for companies like Borland who sell shrinkwrapped boxes with CDs inside. Another reason that the BorCon presentation went over better than the web site materials is that the former mentioned Delphi, whereas the latter does not. People in the Delphi community are sensitive about that. Perhaps overly sensitive, but it’s not something which is worth losing audience members over when you can work around the problem by simply name-dropping your product. An effective message needs to build on things that people already know and like. Delphi 2005 is a great product and gives you lots of ammunition for this. People love the history window and the integrated StarTeam support. Simply saying that Borland intends to do for the other aspects of the business of software what the StarTeam integration in D2005 does for version control and change management will, I think, be a powerful message.
• One powerful way to tell developers that you feel SDO is relevant to them is to have an affordably-priced SKU (thanks to Bob Dawson for pointing this out). Why is there so much more interest in StarTeam in the Delphi community than in CaliberRM? Because StarTeam has an affordable low-end SKU, and CaliberRM doesn’t. Why did Kylix get such a rocky start? Because it was mispriced. Kylix had many other flaws, but the price was the one that people couldn’t work around.
• There’s nothing wrong with having a separate Enterprise conference, as long as Borland doesn’t gut the Developer conference to make it happen. After all, Borland serves different markets, which is why Borland products have many SKUs and sales options. But don’t ever fall into the trap of thinking that business processes are only for the big guys. That’s a mistake that leads to bankruptcy for anyone who makes it.

Update (23 March 2005): Here’s Allen Bauer’s comments on this issue.

Share This | Email this page to a friend

Server Response from: dnrh1.codegear.com