Once you announce a date, you’re already late

Steve Sinofsky (who used to blog on careers at Microsoft for college grads and more recently oversaw the Engineering Windows 7 blog) said something that stuck with me: Once you announce a date, you're already late.

Such is the nature of media coverage of the software industry. You can talk all you want about what the product does, but once you begin to speculate about a ship date, all your discussion may as well never have happened. All anybody cares about any more is how far behind schedule you are. (And as far as the media are concerned, you are already behind schedule.) The ship date countdown clock goes up, and all anybody writes about is how your product has been delayed a week, a month, whatever. Even if your original schedule was a purely speculative guess that you pulled out of your butt and the updated schedule is based on some preliminary analysis, the headline won't be "Microsoft refines schedule for Project X." It'll be "Project X slips three months." The article will probably begin "Project X, announced by Microsoft scarcely two weeks ago, is already in deep trouble, suffering a three-month schedule slip even before it gets out of the gate."

I was reminded of this adage (and use it as my standard answer to questions I'm not authorized to answer) when a Microsoft executive attending RSA Conference 2007, a computer security conference, was asked to speculate on when the next version of Windows would be ready. If you read what he actually said, you can tell that he was just making a guess on the spot, but the next thing you know, the "ship date" makes not only the article headline, but also becomes the top headline on InfoWorld's conference coverage and sparks a discussion on Slashdot. Up goes the countdown clock. Even long after the conference is over, it still is listed as the number two headline for the entire conference, even though it has nothing to do with the conference!

We don't know what it is, but whatever it is, it's already late.

Bonus example:

What started out as a plan turned into a confirmed ship date and then a promise.

Comments (30)
  1. John says:

    On the flip side, if you don’t announce a date then you’ll never actually ship.  RIP Duke Nukem Forever.

  2. Random832 says:

    So is it intentional that this is posted on the (much more recently announced) Windows 7 GA date, months ahead of the "confirmed" January 2010 date?

  3. CGomez says:

    Most of the tech rags, I mean blogs, are stupid.  They just want to drive traffic.  What better way to drive traffic than to take a non-promise and write a headline calling it a promise?  That’s great for SEO and all that matters is traffic.

    Most sites don’t care if they are lying.  All that matters are views.

  4. super secret says:

    A Windows 7 followup is already in the works, is on it’s way. Remember, you read it here first.

  5. James Schend says:

    CGomez: maybe they should pretend they have a kid in a balloon.

  6. Nish says:

    I always thought there were different levels of dates. One that Sales/Marketing wants, one that the Managers want, one that the Developers have in mind, and then at least a couple more, one from the pro-media, and one from the slash-dot-ers. Of course all these dates will be different (to varying degrees) from the eventual release date.

  7. Billy says:

    It does not matter what you say,

    It only matters what they hear. (Or want to hear) ((Or..Or) Think they hear))


  8. Trevel says:

    This is why I try to go ridiculously overboard with my work estimates these days. My coworkers have generally figured out the code (A "four hour" change is one that I’ve finished implementing while they were still talking), but it’s a nice way to manage expectations. And make me look good when I come in way ahead of schedule every time.

  9. Josh says:

    Nish: You forgot the date Testers want, which is always years after the one Devs want :)

  10. McMillan says:

    Looks like Fathi got the release date more than right – maybe there’s a promotion in the works for him.

    But where’s my new user interface paradigm for consumers?

  11. Anonymous Coward says:

    John, there were worse problems in the DNF project than the lack of a release date. You don’t need a release date to ship, except in the sense that once you ship you’ve created a release date. What you do need is a team that is actually working, fixed goals to work toward to prevent feature creep, and a milestone system. There is nothing wrong with the ‘when it’s done, it’s done’ attitude as such. And in cases where it fails, regular release dates would simply slip until the project gets cancelled too.

  12. Nish says:

    Josh : that’s a good catch – can’t believe I missed the QA-date :-)

  13. John says:

    @Anonymous:  I’m not saying "when it’s done" isn’t a valid model, but when you’re working with a deadline it makes doing all the things you mention easier.

  14. Anonymous Coward says:

    John, I tend to agree with you, with the slight proviso that easier isn’t always better.

  15. Miral says:

    This is why people shouldn’t use words like "plan".  To some people, a "plan" is set in stone.

    Something like "estimate" or "guess" might have been better.

  16. Anonymous says:

    "Even if your original schedule was a purely speculative guess that you pulled out of your butt"

    There’s your problem. Its your own damn fault if the media holds you to your word. If you don’t know, don’t say anything. Its called PR, idiot.

    Part of the problem, too, is language. Subtlety just doesn’t work on the international stage. You might say "plan" implying "hope", but someone else can easily and justifiably understand it to mean you’ve got a "contract".

    It’s also a matter of perspective. If you’re waiting for a seat at a popular restaurant, if the waitress says to "expect a 15 to 20 minute wait", you’re going to be pestering her the second you count off 14 minutes. Obviously, though, you know she was only guesstimating and that such things are completely unpredictable.

  17. Matt Green says:

    I hate how people talk about Slashdot as if it has journalistic credibility.

  18. John says:

    Matt Green:  You mean to tell me that kdawson isn’t on the short list of Pulitzer nominees?  I am shocked, shocked!

  19. Worf says:

    I believe one of the lessons I picked up from my Professional Engineer courses was how to deal with the press. The short answer – you don’t, because everything you say will be scrutinized, and everything you didn’t say will be revealed as truth. Pass the question to the communications person.

    It sounds evasive, but that’s the proper way. Not doing so can inadvertently open up legal liability, sow confusion, or most commonly, be misinterpreted. Especially if they want their answer condensed into 10 words, but you really need 100. If you let them condense, it’ll never be good.

    The press know this – which is why they’ll push their microphones on anyone they can, hoping for a non-filtered report. They are no one’s friend, they want a story. And even the blandest answer will give them one.

  20. El Dorko says:

    Same thing happens with prices btw. I’ve been in business for about two decades now, and every time there is a new project it starts out with the customer going "how much will it cost", then me "well I can’t say really, it depends on what it is – what functionality, how much features and so on". Customer: "Yea but roughly, just an estimate". Me: "I don’t want to say anything to that, because you’ll hold me to it in the end". Customer "No I won’t, really, I won’t". Me: "<sigh>ok, maybe something like 60.000".

    Months later, the project morphed into something completely different than what was initially discussed, the customer goes "But you said it would only cost us 60k – why is it now 90k?".

    Lesson to be learned: don’t give out estimates. They’re just not worth anything.

  21. Aaargh! says:

    There is something to be said for a certain fruit company’s strategy on announcing products. "Oh, one more thing … <some new product announced> … and it’s available tomorrow at your local greengrocer"

    You don’t get the whole slipping-date problem, and it also allows you to completely cancel products that turn out not to be as good as you thought they would be. Plus, free marketing due to al the speculation on rumor sites.

  22. Les says:

    It is the reporter’s job to generate revenue for his/her employer.  They are like the used car salesman in the plaid suit.  Twist things, even lie a bit to make the story sound better.  When a reporter reports on another reporters report, hang on, it’s going to be a wild ride.

  23. Evan says:


    Taking lessons from a certain chief engineer of the Enterprise are we?

  24. mikeb says:

    @Michael Grier:  great little article!  I love the complaint that "to add insult to injury no new date was given" when no date was ever given in the first place…

  25. Marcus says:


    "Announce it when it’s done" only works if you have no partners.  Otherwise, either your partners are spluttering and foaming that you didn’t give them any notice to prepare their products to work with yours or you tell your partners and they leak the information anyway.

  26. someone else says:

    Confuse them with unspecific answers long enough to get away, like:

    * When the cows come home.

    * Right after I change into my fur bikini.

    * When I’m done with your mom.

  27. Aaargh! says:


    That’s what NDA’s are for. The iFruit company also works with partners, and some information always leaks but that only fuels the rumors (= good marketing), nothing is officially announced.

  28. someone else says:

    And just so you know, Windows 8 will be released in 2012, including a 128-bit version.

    Wikipedia sez so.

  29. xor says:

    I hate how people talk about Slashdot as if it has journalistic credibility.


Comments are closed.

Skip to main content