Conference 7.286::digital

    1) provide a written guarantee to customers of the product that it will
    be sold and serviced for a certain number of years. This is quite
    common in other industries. If I buy Royal Doulton china, then for some
    (not all) of their patterns I get a guarantee that I will be able to
    buy replacements for broken plates, etc. for the next 10 years with the
    same style and pattern.
    
    This guarantee is not just a guarantee to customers, giving them
    confidence in the lifetime of a product, it is also a guarantee for
    employees. It says to an employee "In x years time we will still be
    selling and supporting this product, and if, at that time you are the
    best person available to do the job you will have a job". Some of us
    who have mortgages and families are motivated by job security as much
    as working on the latest new product. Working on a new product gives
    *no* job security. Remember the PDP-16, PDP-11/60, Trax, ...?  The only
    reason that some of DEC's employees  that became experts on those
    systems still have jobs is that DEC was rapidly expanding at the time
    and could easily move them. Any new product can be a failure.
    
    2) Convince the employees and customers that the above customer 
    commitment is binding. Doulton has kept this type of commitment over at
    least 50 years or so, so you can start to believe them.

    re .1 the markets aren't the same-you can do this with Royal Doulton
    because it never goes out of fashion and doesn't require massive
    capital to reengineer.We used to do just this with Vaxes until quiter
    recently and are constrained by law in some countries to ensure 10
    years supply-however,since it is a computer we supply we know it is
    unlikely to be needed.
    Sorry,but the only way to do rampdowns is not to dither around-decide
    and kill instantly has always been the strategy of successful
    companies.It is painful but it is like pulling teeth as apposed to six
    months toothache.
    I also believe it is quite refreshing to do-you avoid a lot of the
    uncertainty problems with employees and customers and if you deal with
    clients directly it is often painless and manageable.
    Our problem,as in most domains,is management by rumor and dithering.
    Now can I be a VP please?

    	I agree that some things go out of fashion, but they can take a
    long time, even in the computer business. If DEC had produced a
    portable Fortran compiler at the time it produced the PDP-6 we might be
    celebrating 30 years of very boring, very profitable updates, upgrades,
    simple ports to new platforms, and DEC would have been a little richer.
    If Unix had gone the way I thought it ought to go when I first saw it
    then it would have been dead 20 years ago, but maybe it still has a
    year or two left of profitability, even if it is ancient and boring.
    
    	What customers want is stability. This is the reason for the
    popularity of open systems. The customer *knows* that some of his
    suppliers will go bankrupt, change their line of business, or whatever
    in the next few years, and he wants some sort of guarantee for the
    product he is interested in. Where we have a unique and differentiating
    product (knowing the basenoter I would guess we might be talking about
    VTX here) then possibly we don't want to ensure that we have
    competition by showing the customer an alternative supplier.
    
    	And it is the *old* products that make the profit. At the time DEC
    was writing VMS, if we had scrapped *all* our PDP-11 products we would
    still have been more profitable than DG or Prime because of our PDP-8
    business. If we had tried to kill that quickly as "obselete" then we
    wouldn't have had the money to develop and market the VAX, which in
    turn didn't make much money in its first 5 years of business. We still
    have a significant PDP-11 business, after 2 years of AXP.

    re .3
    
    Excellent point. We should be focussing on products (old, or even if
    they are new) that are not profitable, and killing them quickly.
    
    Even if something is old (and perhaps rather inelegant by modern
    standards), if there is a profitable market, we should retain (even
    enhance ?) it. Note that IBM will still sell you VSE on mainframes (why ?
    because they can make money - even though Fred and Wilmah would be
    familiar with it.)
    
    I suppose that then raises another question - do we *know* how
    profitable any particular product/offering may be ?
    
    Oh, and the quickest way to 'dump' a product ? - outsource it, plus any
    staff that need to go with it. There will always be a player who can
    produce a return to suit his requirements, even with 'old' products.
    That way, Digital could still support the installed base, via a
    'partner' instead. 
    
    
    AW