Friday, February 10, 2006

Exposing the "Cash Cow" Myth.

“Cash Cow.”  Wikipedia defines it like this: “In business, a cash cow is a product or a business unit that generates unusually high profit margins: so high that it is responsible for a large amount of a company's operating profit. This profit far exceeds the amount necessary to maintain the cash cow business, and the excess is used by the business for other purposes. The expression is a metaphor for a dairy cow, which after being acquired can be milked on an ongoing basis with little expense.”

When this unfortunate term was applied to Delphi by a former interim CEO, I've seen it being parroted extensively in newsgroup postings and blog content.  While having a cash cow business is usually a good thing for a company that is looking to diversify and move into other markets, it is also a very dangerous position to be in for that cash cow business.  According to the above definition, “This profit far exceeds the amount necessary to maintain the cash cow business, and the excess is used by the business for other purposes.”  Let us examine whether or not this applies to Delphi;  It has been no secret that Delphi 2005 was... ahem... a little less than what would be considered a quality release.  If you examine the events that lead up to that release, it was clear that the management at that time actually did come to the false conclusion that Borland's developer tools were a “cash cow” business.

If the business were truly a cash cow, then minimal investment and commitment would have been all that is needed to maintain that stead cash flow.  Instead the opposite happened.  The uproar among the community was felt throughout all facets of Borland.  This lead to a renewed commitment to making sure the team was given the appropriate attention and resources to ensure that Delphi 2006 would not be a repeat performance.  By all accounts, that change in how management funded the project had a profound affect on D2006.  Which tends to explode the cash-cow myth because as soon as focus was shifted away the effects were immediate.  Yet, the “cash cow” myth seemed to remain.

Enter Tod Nielsen.  One thing that truly excited me about Tod was that he really understood what kind of investment was needed to build a complete, end-to-end development tool.  If you think about it, Delphi and C++Builder start from essentially zero.  We control the compiler, tools, frameworks, IDE, debuggers, etc...  There isn't an established compiler and framework around which tooling is built.  It is built from the ground up.  Kind of like heading out to the iron ore mine with the intent to build a car.  Other tools merely need to visit the machine shop.  I discussed this difference in detail in this posting.

Now this is pure speculation, but from my vantage point this is what I think Tod encountered as he began to assess the current state of affairs here at Borland.  On one had he came in and saw a strong legacy of popular development tools with a strong, loyal customer base.  On the other he saw a strong suite of tools that support management of the application lifecycle.  He also saw a huge scism between the way these two sets of tools are marketed, delivered, and sold.  With the dev tools, there is a strong, what is called “channel presence.”  This is a more direct to the developer model.  That is clearly where the “IDE” products thrive.  Then there is all the other things which require a long-lead, C-Level, VP of AppDev, consultancy, schmoozing kind of sale.  This is where tools such as CaliberRM, StarTeam, Tempo, (and now Segue) live.

Rather than being blinded by the new bright shiny object that is ALM/SDO, he began to take a more holistic and pragmatic approach.  He knows that in order for both to survive they needed a high degree of focus, investment, and commitment.  The problem is that Borland is in this constant tug-of-war.  It is easy to be on one side or the other and simply say, “Well just invest in both!  How hard can that be?”  The problem is, and I've said this many times before, that Borland is simply too big to act small, and too small to act big.  Tod, knowing that Borland just cannot do both things at the same time and with the same focus, he and the executive team took this radical approach.

So that is where we are now;  On the cusp of a new era.  There's sure to be a bumpy road ahead, but it's going to exciting and I truly believe, will turn out to be the best thing that could have happened to Borland and whatever the new dev tools company ends up being.  As an aside, I remember many years ago sitting in Gary Whizin's office (the then long time Turbo Pascal/Delphi R&D manager), with Chuck Jazdzewki and Anders Hejlsberg, potificating and fantasizing about the prospect of spinning off the Delphi product into a private company.  So this is not really a new idea, merely an idea whose time has finally come.

I encourage you to reserve judgment until the dust settles and all the i's are dotted and t's are crossed.  In fact, things are moving full steam ahead with our current roadmap, a new Delphi 2006 trial edition is in the works as is some new product updates and some other things I'm sure the community will be excited about.  So for the Delphi/C++Builder team, it is business as usual.


  1. Wow, that attitude shakeup I suggested in other posts is already starting to happen. Did you actually suggest that D2K5 was, perhaps, FLAWED? A position you openly critized me for taking in one of your earlier blog posts, in a tone which seemed to imply I had no right to complain? Glad to see yer on the bus, because only then can real progress actually happen. Let me know when the update is ready, I'll check it out.

    As for "Cashcow", we that is minimal investment for massive return. It does not indicate scale, but rather RELATIVE scale. Certainly most companies did not few a extra hundred million dollars lying around extra to buy other companies with. The reason people consider Delphi to be a cash cow is that its roi is clearly EXTRODINARILLY high compared to say books, clothing, food, cars, houses... The fact that management tends to fritter that return away on boondogles doesn't change the exceptionally high quality of the return.

    Sadly, the high quality of the return may just explain the management boondogles themselves.

    And yes, Delphi *IS* expensive to develop, after all, people don't work for free, offices aren't cheap and power isn't generated by entertaining chimps on rows of bicycles hooked to generators (and even if they were, that many bananas would be almost as expensive as the uh "cartage") As a result, the sold off (not spun off, they are talking about SELLING Delphi, not spinning it off) unit would need an owner not just with deep enough pockets to buy it but to also fund it's ongoing development for at least 3 years before a return on investment should be considered adequate for self sustanence.

    If anyone there has any MS prejudices, you might wanna start working the rough edges off them, because the numbers required are HIGHLY suggestive.

  2. Well, thanks for the interesting blog. I've been a long time borland user, but have started wondering why I continually buy the product. I'm seriously considering switching to Visual Studio (already have for mobile development I can't wait on you guys). And I'm not finding it that bad. Esp when Delphi and VS are looking more and more alike.

    I seriously hope that "Turbo" or "New Borland" or whatever your new company is pulls it off. My choices have become more and more limited over the years, and I'm down to 3. MS,Borland or GNU. I would hate to think that the only way to compete with MS is FOSS.

  3. Would the new company consider using ALM for Delphi and Kylix?

    I think if the people, processes and technology required to maximize the business value of Delphi and Kylix were aligned we could see improved results. It would be unfortunate if these projects became another statistic on the Chaos Casualty report.

    -Bill Smith

  4. I guess we're all a little scared of it being MS because they have a habbit of taking a product and either it goes in a direction people didnt really want (see visual foxpro what a waste) or ceases to be, or, we just fear they wont love it half as much as we do.

    Thanks for a great blog Allen.

  5. MS isn't going to buy the IDE tools group - why would they? The hassle to try to integrate the products, and provide a reasonable story going forward would far exceed any benefits possible from an acquisition.

    MS is not in a weak position where they need something like Delphi. They have just completed their most important set of developer tool releases ever in their history after years of heavy investment. The next couple of years will see payback of that investment, as MS works to develop their next release.

    This is what should be going on at Borland, but it seems that the investment didn't happen along the way. Maybe these tools were treated as cash cows, but managment forgot that today's investment is still required for tomorrow's returns - the law of the harvest. I think the IDE tools could and should be a cash cow - but clearly ongoing investment is always needed. This should come as a surprise to nobody, however!

  6. About Gary, is he still working as a teacher?

  7. I believe there is a real possability that MS or other large corporation would be very intersted inthe IDE division. But not for the products.

    Possibly the most valuable assest the IDE division has is it's patents. I seems to recall a settlement just afew years ago where MS bought $100M in Borland because of patent issues, do they really want a competitor buying those patents? Sure they have aright to use them now, but still ...

    Secondly would be the employees, I would image they could be very valuable for the right buyer.

    Thirdly would be the technology itself, ot rather the knowhow to put that technology into a product effectively.

    I'm not convinced the actual products have much of a home in any of the existing busineses I can think of.

  8. I think that Delphi story is at the end, and Borland is trying to sell it before total crash. No serious company will by Delphi.For developers, Kylix was a sign in what direction things are going, and how future is going to look with MS v.s. Borland. Poor future vision from Borland...Sadly for us...


Please keep your comments related to the post on which you are commenting. No spam, personal attacks, or general nastiness. I will be watching and will delete comments I find irrelevant, offensive and unnecessary.