Conference 7.286::digital

    Go talk to Greg Robert. He was the prime original visionary that
    brought about Software Licensing in DEC. He still has a clear vision of
    what needs to be done, inspite of the years of battling that he has
    done with the DEC bureaucracy.
    
    I'll add my 2� worth later.
    
    Peter Q.
    (Former Licensing Developer, and continuing Software Business
    Evangelist.)
    

    
    For some products we require the customer to order a license (part
    numbe and price) and themedia/docs (separate part number and price).
    It doesn't make any sense to require a customer to buy two things
    which in his mind are only one thing. 
    
    That should be your principle: If a customer is buying a single
    product, there should be one part number, and one price. If this
    means lots of variations, okay. But the customer only has to go
    down a list and pick the variant he wants/needs, then orders it. 
    

    I don't think the part numbering scheme is the problem per see.
    Changing the way part numbers are done will only attack the symptom of
    the problem not the problem itself.
    
    At SOME level you need a way to differentiate between multiple options
    eg tiers, processors, language etc.
    
    The problem is that at present the PERSON doing the quote has to
    construct the complicated part numbers. SOFTWARE should be doing this.
    What's needed is smart software with very friendly menu driven
    interfaces that ask the questions the customer would ask and then that
    would construct the part number. The part number should only be seen
    by internal Digital systems (note I said systems not people). In that
    way you can come up with a really complex part number that gives you
    all the flexibility you like.
    
    It really annoys me to hear people say:
    
    	"The problem is that there are too many part numbers"
    
    WRONG
    
    	"The problem is that there isn't a decent AI type system that can
    	 translate the language of the salesrep and the customer into
    	 a set of part numbers".
    
    Dave

Dean,

Good luck in fixing this problem.  The first thing I would do if I were you 
is go to Fry's or COMPUSA or some other super store and see how they can 
make literally thousands of computer parts available without any part 
numbers at all from the end user perspective.  Notice that all other 
software is available with media and license packaged together?  This is a 
good thing.  You may also notice that there is no difference in price or 
licensing when changing hardware, (i.e. Compaq with 5 1/4" versus Zeos with 
3 1/2" floppy.  Nor does it matter whether the user has a 286, 386, 486, 
new pentium seed, or V20 based system.  If it runs the OS, the license and 
media are not differentiated.

From the end user perspective, if you change hardware without changing the 
number of users, then the software company may only know if you request a 
media change.  Digital should do the same.  I have custis that use Suns, 
HPs, RS6000s and DEC stuff.  I do not want separate part numbers for 
licenses and support.  A simple part name and media request should suffice. 
Perhaps a simple Yes/No check sheet for desired support.  (No installation, 
Yes Media and Doc updates...)

It is currently impossible for my standard custi who changes HW and 
locations constantly to stay legal with our licensing scheme.  We supply 
waivers for our custis that say as long as they stay under the allowed 
number of licenses, they can use them on any hardware they like.  

As DEC gets better (hopefully) at writing and selling SW on other folks HW, 
we need to get better at making ordering and staying legal easier.


For your questions:

1) I doubt there is any intelligence left in our part numbers, only data 
that a few gurus can decipher.  Ordering equivalent licenses for different 
hardware is very difficult based on the part numbering system.  How can I 
honestly tell an HP9000/400 license from an HP9000/700 license in a network 
model?  Why should I?  I must do so that I do not need a lawyer to write a 
waiver (requiring negotiation) for each custi.

2)  I still have a hard time configuring quotes and updates for custis that 
upgrade from old Sun Sparcs to NEC AWS workstations to HP9000/700s.  I know 
I have given out illegal quotes simply because the Product Manager, the CAS 
people and I could not determine the proper part numbers for hardware that 
the part numbering system system does not recognize.  (NEC running Kanji in 
the U.S....)

I have spent literally hours creating quotes that had 80 part numbers where 
we were trying to sell only 4 things!  Two types of licenses, support, and 
media for each of 6 hardware/OS types.  WE ended up creating an English 
language versin of the quote that we all use as the version we are holding 
each other to as nobody is quite sure the AQS version is accurate.  I find 
this difficult to explain to custis that I am trying to convince should buy 
automation consulting from DEC.

3)  I doubt processor tier has any future need to exist.  It causes people 
to create suboptimal systems in order to live within our pricing scheme.  I 
think some users like them only because it is a necessary evil tht they 
understand.  Can anybody explain to me the difference between a license 
type and license rating?

How about a simple system that looks like a matrix for each solution being 
purchased.

		Qnty.	Lic type	Ship date	Vers.	Other attr.
SW name
Media
Support
Updates
OS


Where required layered products and other requirements fall out of the 
quote.

I think having a system that get more complex will only cause more 
technical violations.  I would bet we have minimized these to date simply 
because of our collective inability to create and sell software that is 
successful on non DEC boxes.  As we get better ( and if we do not it won't 
matter what our part numbering scheme is) at writing products the problems 
will only get worse.

matt

    1. Make it simpler.
    
    VTX price on QL100 brings up 48 variants of Fortran (QL100) licensing.
    This doesn't include the QA100 and other variants of media support etc. 
    Each software product must have 200 part numbers associated with it,
    which can be confusing.
    
    2. Standardize groups.
    
    A personal use license QL-<upi>AA-2B for C++ costs $2000, for C $1120,
    for Basic $1030, but for ADA $8150. INtroduce the concept of buying a
    "language" license at a standard price.
    
    
    q
    

    As a sales rep selling the software is easy.  The hard part is quoting
    it.
    
    The folks in the EDI Practice made it easy for me.  One part number
    gets me the required software license, media, documentation,
    installation, updates and installation.
    
    The installation required sending an EDI specialist out here to do the
    install and test verification.  The one week he spent here resulted in
    over $200,000 of additional consultation and implemenmtation services.
    
    I wish all the software had the services built in like this.  besides,
    if we can get one of our consultants on site for a week , selling their
    services is a lot easier.
    
    Actually, there are more than one part number.  The variants were the
    one digit power rating and a one digit identifier to the communication
    component.
    
    Ron
    

re              <<< Note 2196.6 by WR1FOR::LEZAMA_RO >>>
>               -< easier software quotes >-


WOW!  Every product manager could learn from that lesson. It sounds like 
that is a model to consider.

matt

.6> As a sales rep selling the software is easy.  The hard part is quoting
.6> it.

Amen.  I was surprised to hear the word "strategy" in the base note.

I don't/won't even come close to trying to understand software licensing
and part numbers.  To do so would be to take an enormous amount of otherwise
productive selling time and put it into ... paperwork.

I gave up on quoting software a long time ago.  Now, I call 1-800-DECSALE
and get licensing people.  They are *great*.  Wonderful people.  In about
5 minutes I can accomplish what would probably take an hour or better just
to get it wrong .. if I tried to do it "by the book".

Bubba

    I agree with previous notes.  Simplify, Simplify, Simplify.  I, too,
    have spent hours on quotes trying to get the software part numbers
    correct.  This is after the sales person has said he can't figure it
    out.  I'm in sales support and the last thing I want to do is spend
    time trying to figure out part numbers for software.  But, to make the
    sale, I work on it.  We recently had one quote where Sales, Sales
    Support, Services Support, and the hotline were trying to come up with
    the proper part numbers.  All told we must have spent a totals of about
    10-12 person hours.  And this was on an upgrade from one Digital
    processor to another!
    
    The idea of the quoting software taking care of the `part number' is
    about the best, if you must retain some kind of complexity.  But I
    think that the best way is to come up with a unit or per user price
    that gets lower as a higher volume is purchased.  Let the system
    automatically compute the discount; don't add another part number.  Who
    cares what processor it is on.  If only one person needs to use it,
    charge them for one user.  If ten need to use it, charge them for ten. 
    If five are on Sun and five are on Digital, who cares.  This is true
    commodity pricing.  Does Microsoft care that MS C is running on a
    DELL PC or a Digital PC?  Of course not.  Yet everyone has to purchase
    a copy of MS C if they want to run it.  If I run it on a 286 or a 486,
    it costs me the exact same amount to purchase.
    
    Obviously, there has to be a way to specify the vendor's hardware so
    that the proper binaries get shipped.  And a way to specify the proper
    media.  And documentation, if they want it.  But these should all be
    figured out by the software.  The sales rep should not have to try to
    construct the part number. 
    
    
    I could go on and on about this, because it really is a headache.
    
    tgc

Re .5:

>    2. Standardize groups.
>
>    A personal use license QL-<upi>AA-2B for C++ costs $2000, for C $1120,
>    for Basic $1030, but for ADA $8150. INtroduce the concept of buying a
>    "language" license at a standard price.

Yeah, I hear GM is going to a standard price for cars.  Chevette, Caddy,
whatever - one price for any car on the lot.  The salesmen are a lot happier,
because this is much more convenient for them.
				/AHM

    
    Check out the Dell catalog. Not one part number. 
    
    

    Simply ask yourself this question:  "Is there anything about our
    current part numbering system that adds value FOR THE CUSTOMER?"
    
    \dave

    two points in regards to part numbers:
    
    I know of at least one state that dropped a 16 digit "smart" number
    that contained all types of embedded information for a 9 digit number
    (yes it is SSN but please don't go down that rathole). In all cases,
    the DMV and police could get the info they needed much more efficiently
    from the computer system. At the time that computer access was
    expensive or difficult, it made sense. It doesn't any more.
    
    Second. A non "smart" internal identifier that the customer and/or
    sales rep rarely sees should be self validating. Use a checksum
    methodoly, not a look up in the master table methodology. It will make
    the next step of getting the "catalog" into the field on notebook pcs
    significantly eaiser and more efficient. Look at the hand help scanner
    in any supper market.
    
    Scot.

    The Dell approach is a good model, in a minimal sort of way.
    Certainly they must have part numbers, but apparently they don't
    impose them on their customers.  We shouldn't either.

    Ultimately our problem is (and has been for years) that we insist on
    trying to pack all manner of information into a part identifier, so that
    the identifier by itself is a meaningful, portable, convenient repository
    of information as long as one understands increasingly complex structure
    rules.  If you know how to parse the identifier you could determine the
    media for a software product, or the warranty terms, or the quantity,
    or the screen color...

    Here's the new 'minimal' product identifier: number & suffix.
    The 'number' is an arbitrary and random.  The 'suffix' is also a number
    used to identify variations for logistical purposes.  No meaning is to
    be directly associated with either.  For all the interesting & important
    items of information about a product, use the product identifer and consult
    the (universally accessible!) products database.
    
    It could happen!

This is an obvious problem for a cross-functional team.  Put together a 
team including representatives from:

	Corporate Business Practices
	Software Product Management (possibly more than one)
	Sales
	Marketing
	DEC direct/DesktopDirect (possibly both)
	SSB
	International Software Product Management/Marketing (maybe)
	Customers (maybe)
	Anyone else?

Actually, it isn't clear whether the Business Practices group has any stake 
in this, other than as maintainers of the standard.  At best, they're a 
supplier of the mechanism, not a user.  Six to eight people seems like a 
manageable team.

This is also an area that's amenable to QFD analysis and Contextual 
Inquiry, or similar tools for collecting and analyzing customer benefits.

   Gary

Here is my opinion about how we should think about this area.
    

Scenario:

   Customer wants the capability of doing X information processing.
   Customer may already have a computer (or computers) for which a
   program called A which does X is available.  A Digital employee
   wishes to record one or more facts associated with a transaction 
   to sell a copy of A to the customer.
   
Question:

   What kind of reference information is needed for such a transaction
   and its associated business activities?
   
Answer:

   1) Describe the customer and his existing system.
   2) Describe X.
   3) Identify A as a program which does X.
   4) Determine what the customer needs to buy besides A.
   5) Determine the price to be quoted for 4).
   6) Deliver the price information in 5) to the customer.
   7) Negotiate the sale, using whatever leverage is available.
   8) If negotiation succeeds then go to step 9). 
      Otherwise close the transaction with appropriate failure
      analysis.  
   9) Update the proposal with the negotiated price.  
  10) Trigger the execution of the proposal as updated in 8).
   

What does this have to do with a part number for software?

   Well, program A is not really just one thing, but a class of
   things that we refer to with one name to avoid very long
   descriptions.  So the essential function is to translate X
   into A modified by all the relevant characteristics of the
   customer's system to make sure that A really serves the need.
   We don't need to identify all the customer variations that
   program A can exist in until we have such a customer.  Our
   solution will be to have one part number per existing
   application, regardless of the customer specifications.
   
   We usually translate X into A by a human's (salesperson)
   interpretation of what the customer says X is.  But that
   interpretation could be mechanized with some investment.
   
   Second, we need to record some critical data about the
   customer in the customer description that will enable us to
   select that specific A that will meet the customer's needs.
   
   So, the solution is to 
   
      a) design a customer description system that will describe
         the customer's computer configuration(s) adequately and
      
      b) design a solution description system that requests
         configuration data from the customer file and either
         
      c) selects a pre-configured program or 
      
      d) configures a program 
      
   for sale to the customer.  The existence of a part number
   beyond that which identifies A is unnecessary.  We do it
   today because we have a build-to-forecast mentality in our
   system design.  We need to develop the engineer-to-order
   mentality used in aerospace and other long range project
   businesses.
   
   Incidentally, note that this obsoletes many "software
   protection solutions" by providing the customer with a
   specially configured solution which integrates seamlessly
   into the customer's environment, perhaps even delivered
   electronically.
   
Dick
    

    All that has been said sounds good.  The K.I.S.S. should be focused
    on.  When developing these changes please include the Customers,
    especially when looking into how to handle the services.  Many of them
    have dealt with our competition and could probably shed some light on
    how it can better be handled.  If we are to "Delight our Customer" we
    need to include them in helping us design the ways we do business.
    
    IMHO
    Tom

The root of the problem is in our business practices. We're far from best
in class from a simplicity viewpoint. I have four people in SW licencing
here; Microsoft have 1 and Lotus have 1, who's time is spent in compliance
education rather than "guideline" interpretation.

When it gets to voting, the geographies have limited ability to alter the
course of things. If the US, Europe and GIA agree, they can still *easily* be
outvoted on the appropriate corporate committees today.

Part numbers are but a symptom of the results!
								- Ian W.

    This is probably the right notes file to solicit inputs to your
    questions. At some point in time the base note and replys should be
    crossposted to the Software Licensing Notes file so that management
    will be exposed to the issues.
    
    The Software Licensing Notes File is located at:
    
                     YUPPIE::SOFTWARE_LICENSING
    
    John

    Think about how you would construct a solution approach with the
    customer...  After you have understood the problem to a sufficient
    extent then you begin to build a configuration.  You might use a
    white board or a flip chart or just sketch on a piece of paper.
    You draw block diagrams - a VAX here, a PC there, a workstation
    over here, string some Ethernet, maybe a few terminal servers,
    printers, modems.  Then you configure the peripherals - disks and
    controllers, tapes, etc.  You configure the software similarly -
    compilers and CASE tools here, forms manager there, database on
    this node, spreadsheets and word processors over there.  Now you
    want a �5% price on that configuration.  Now you take another look
    at the system and play `What if' - What if I moved the front-end
    down to PCs?  What if I add a second VAX and share the disks?
    What if I need another network protocol?  Now price it again.
    Iterate a couple more times.  Then leave sufficient information
    behind so that the customer can continue to play a few more `What
    if' scenarios after you leave as new ideas occur to him/her.

    Total elapsed time from sketch completion to pricing for each
    permutation should be 10-15 minutes maximum.  If your
    part-number/licensing/quoting scheme doesn't pass this test, then
    try again.

	Mike

    Just for everyone's information, there is a configuration tool today,
    that is easy to use and fast, that will configure software part numbers
    for licenses and media (H-kits).  It's called the Software License
    Configurator (SLC).  It is available in AQS, the Account Workbench, and
    the Electronic Connection.  It can be accessed in AQS, either directly
    from the quote (GOLD/C) or from the AQS Main Menu (Option 14).  It can
    be accessed in the Account Workbench under the Configuration Tool
    menu.  It can be accessed in the Electronic Connection under the
    Software category and then the Licensing category.
    
    Regards!
    
    Steve
      

RE: .0

I have a question...

Wht will be done with this report?

Will anyone with authority read it?

Has anyone with authority signed up to act on it?

By authority I mean someone that can, by themself, make a change.

The only person that I am aware of that can do this is Bob Palmer.

About one year ago, I embarked on a crusade to simplify software 
licensing.

My crusade ended up in front of the Software Business Practices Committee
last may.

I said my piece, was thanked and congratulated for a good job...AND 
NOTHING  HAS HAPPENED SINCE!!!

The people on that committee all report to VPs.  They told me that they 
could not actually make any changes.  One told me that I had more power 
than he did.

In general, the people on the committee knew about most of the issues I
brought up.  The problem is that they could not act on any of them.

I have since seen the software licensing stuff for Alpha.  It is
obvious to me that nothing has changed.

Many of these notes have good suggestions...

I think that many people know exactly what to do...  For some reason
the corporation just can't get out of the way.

Until someone with a very big club is ready to knock down some stovepipes
we (Digital) are doomed to write a report on this every year with no
fixes

Ken Slattery

    [Synopsis]
    
    So far, almost all of the replies in this note have concentrated on
    ways to make it simpler for a sales rep to figure out what a customer
    wants and then produce a quote.  These methods would have been great
    four years ago, but the times have changed.
    
    We need to make it possible for the customer to *once again* be able
    to pick up a catalog, understand what are software does and what it
    costs, and be able to order it from us in an efficient manner.  If
    we just come up with a new system that masks the complexity to the
    customer, but preserves the overhead internally, then we lose.  Our
    software will be too expensive, or we will lose money on it.
    
    [The verbage]
    
    Our software is not only difficult to configure, quote, and order,
    but it's also expensive.  Most customers want to stay in compliance,
    but our current system makes them spend so much time wrestling with
    licenses that they either give up or get furious.  Most of them have
    realized by now that the DEC rigamarole is a major reason why the
    software costs so much, not because there's any special quality 
    inherent in the product itself.  And computer customers are becoming
    intolerant of vendors who are perceived to be inefficient in any form.
    
    The current system also guarantees that no customer can have any idea
    how much software will cost until both the customer and the sales rep
    have put some major effort into configuring it.  This can also make
    the process very long drawn out if the customer does not have ready
    access to a sales rep (as is very much the case where the *real* 
    customer is an application user, and the sales rep only calls on the
    system manager).
    
    We can mask the problem by using any or all of the following:
    
    	- Remote support (1-800-DECSELL and the like)
    	- Software configuration programs (AQS and SLC)
    	- A lookup-reference PN scheme as opposed to "smart" PNs
    	- Eliminating "cluster" licensing in favor of volume licensing.
    
    But to solve the long-term business problems, I believe that we must
    change both the licensing model and the sales model.  To wit:
    
    	- User or concurrent-based licensing, independent of platform
    	- Keep the PAK option.  Most major PC vendors are now using it
    	- Ship software with pared-down documentation
    	- Discontinue umpteen-jillion media type support
    	- Incentives to migrate customers to CD-ROM
    	- Simplify PN's accordingly and put out a usuable catalog
    
    These moves will reduce our internal expense to sell software, allow
    us to more competitively price our products, and re-empower the
    customer to order and manage software effectively.
    
    Geoff Unland in Austin

Re: .22

Ken,

Can I have a copy of your presentation/proposal please? Lack of empowerment
is a cop-out. I will commit to getting the European Software Business Mgmnt
Team to see/comment on your proposals and then to push.

On the simplicity lark, in my last job (I managed DECdirect UK's Software
capability from birth to $100M+ margin), we processed about 64% of the
total subsidiaries software business through a fax machine.

Having a simple catalogue does help!
								- Ian W.

    If anyone wants to review the state of this discussion as it existed
    about 1 year ago refer to:
    
    Note 1716 in the ASIMOV::MARKETING notes file
    
    This note started as a result of personal licensing and is the start
    of the journey that ended me up at the Business Practices Committee
    that I refered to in an earlier note.
    
    Happy reading                       
    
    Ken Slattery

        One from the base noter...

        Many thanks to those who have responded.  The input clearly shows 
    that we could benefit from a new approach to the problem.  It is 
    obvious that there are many fantastic ideas for what that new approach
    should be.

        In the time that I have been investigating this issue, the
    overhead and difficulties that result from our use of part numbers
    was easy to see.  A few weeks ago, I presented the status of this
    work to the Corporate Business Practices Management Team.  In that
    presentation, I identified the risks as I see them.  ie- 

            -   that the dependencies that we place on part 
                numbers force us to use very complex processes

            -   that the resulting complexity is thrust upon our 
                customers (here I went so far as to bring copies of 
                four pages from a DEC Direct catalog that contained
                exhaustive and confusing instructions on how to
                configure the part number for a license.)

            -   that this complexity will increase at an exponential
                rate once we move forward with the proposed business
                practices of the future (given that we maintain the
                status quo of todays processes)

            -   that if we maintain the status quo, something is going
                to "break".


        Although the mood around the room was basic agreement with what
    was said, the audience didn't feel as though the argument was compelling
    enough to warrant asking the business to invest in making this change.
    They requested that I continue my analysis and QUANTIFY the problem.
    Thus we would be working with a dollars-to-dollars comparison, with a
    clear ROI on the investment.  Fair enough.

        This lead me to post the base note.  What I need is to quantify the
    problem.  I need some real examples of where and how our processes have 
    cost us money.  Have we lost any sales?  Have we pissed off any customers?
    Does anyone agree with my assumption that something will break?  If so, 
    what is that "something"?  Perhaps this problem is too subjective for 
    this type of quantification...?

        Thanks,


                                - Dean -

    Dean,
    
    Lost sales and such are very difficult to use as evidence against the
    current system.  Focus instead on the maintenance cost of the current
    system.  Take the cost of APICS.  Add in the cost of the IT
    organization to support it.  Include the cost of the engineering
    expediters and project leaders who are needed to discipline the process
    of adding part numbers.  Add lost sales due to missing the market
    window due to delays due to complex administration.  If you can't get
    dollar estimates, use headcount as a surrogate.
    
    Another approach is to just set some performance standards for the
    system.  Things like time delays for basic customer related activities,
    responding to an RFQ, developing a proposal, answering a customer
    question in the catalog sales operation, calculating the price of an
    order, etc.  Its hard for people to argue with aggressive performance
    standards.  Show the statistics for today and ask "how good should this
    be?"
    
    Once you have agreement on the standards, challenge the
    supporters of the status quo to show how they can meet the standards. 
    If they fail, you've got evidence that the current system is
    inadequate.  Then you can drive them to do their jobs!
    
    Actually, I would recommend you do both.  Try each out on one member of
    the committee that you can consider an ally.  Get that person's
    rcommendation as to which approach to use first.  Then go for it!  The
    battle will be tough, but it has to be fought if this company is ever
    going to get "lean and mean".
    
    Good luck,
    
    Dick

>     cost us money.  Have we lost any sales?  Have we pissed off any 
customers?

While those are good questions, it's probably easier to start with basics:
How much time does a product manager currently put into dealing with all the 
part numbers and license issues?  How much will a better licensing scheme 
improve that?  Multiply by all the product managers, and their average 
salaries, and you have one chunk of savings.  

Do the same for sales people.  That's another chunk of savings.

Then ask how many sales people don't put the effort into selling software 
because it's too much effort -- take an anonymous poll, if necessary.  
Translate that into lost sales.

Then ask how many new ideas don't get proper review, because product managers 
are too busy dealing with part numbers.  Those are more lost opportunities.

How much does all this complexity cost the SSB in ordinary maintenance?  In 
correcting misships?  That's more savings.

The specific issue of annoying customers and losing sales as a direct result 
of the licensing issues is just a small piece of the pie.  Can the Corporate 
Business Practices Team tell you the total cost, today, of the current system?
If not, it doesn't speak too highly for the current system.

   Gary

re: .27, .28

It's commonplace to get notes collision on simple questions.  When you get 
two notes colliding such as these, which are in basic agreement on a 
complex issue, it says something stronger.

   Gary

    The following comments appear in an otherwise favorable review of
    Vivace from the Nov '92 issue of "Computer Shopper".
    
    	"Eventually we established that Vivace was, in fact, product number
    	QB-MQGAA-SA.1.0 and would run on PCs running windows.  If you call
    	to purchase Viace, keep that number handy.  It'll save you time and
    	aggravation."
    
    Really motivates you to call and buy the product, don't it?
    
    \dave

    1.  When "selling" software I don't "sell" as much as I clarify.  I spend
    MORE TIME explaining and clarifying how to order it than I do selling
    it.  The product information, demos, and sales calls are the easy
    part.  
    
    2.  Some customers just don't get it.  But I don't blame them.  And then 
    they get frustrated and upset.  Just yesterday a customer ordered the
    wrong VAC C license because of our confusing system.  What does that
    cost in Return Authorizations in order to keep customers happy?  In
    terms of contract modifications?  
    
    3.  I've been in internal meetings where we have had four people working
    (arguing) on what the correct S/W part numbers should be.  These people
    were from Sales, Sales Support, Customer Services, and even a Regional
    License Specialist.  Then when we would agree (guess), AQS would tell us
    to piss off.
    
    4.  The DECspeak acronyms (sp) around ordering software are misleading, 
    causing delays in ordering and higher selling costs.  Why LPS instead
    of Phone Service?  What is the real difference of SSS, SNS and SDS anyway?
    And why can't the descriptions spit out by AQS be in English, instead
    of DECspeak	("VAX FMS Lic w/ WARR")????????????  Does anybody know the
    cost of translating AQS to English?  FLAME ON:  Does anybody care?
    
    5.  If we fixed the above problems we would increase sales
    productivity, customer satisfaction, and contribute to the bottom line.
    (How you quantify that is another note, but you could easily calculate
    the cost of four $20+ per hour people doing the same necessary, but
    stupid task.  If the process was simpler, a sales rep wouldn't have
    to call 1-800-DEC-SALE [as mentioned in another note] and Digital could 
    eliminate those support positions.  Ask yourself WHY WE NEED license 
    specialists and DEC-SALE. Eliminate the complexity and you can
    eliminate the positions and eliminate the costs.  Count up how many
    of these people we have.  The goal should be to have a system that
    a customer/secretary/grandmother could use to order DEC software, not
    a highly trained and costly resource.  Heck, you could even eliminate
    the cost of publishing a 979 page U.S. Software Price List  four times
    a year.  How many copies do we print?  How much does it cost to ship
    those?  (And how many trees do we kill?)
    
    6.  We don't need laptops, workbenches or AI software.  We need simplicity.
    
    Good luck.
     

Re: .-last several

80% of DECdirect's Software business in the UK - which in turn is 80% of the
subs software business - comes in via a fax machine. Why not go look at the
European DECdirect Software Catalogue (now also being deployed in GIA soon)? 
Send me mail and i'll send a copy!

I don't buy the story about our future business practices needing to be even
more complex. For most products, user based plus an ability to split and
reconstitute a volume of them would take us down to two parts per product.

This won't do any favours for the many people inside Digital who get paid to
navigate this complexity. Like C Northcote Parkinson says, the number of
enemy killed is inversely proportional to the number of generals on your
own side ;-)

The other thing that bugs me is that you're asking a wide audience for
feedback when this is *never* solicited from the people who operationally
run the software business. But that's another story.
								- Ian W.

    I was going to say I was flabergasted by the reply you got from the
    committee.
    
    On firther thought I've changed my mind. I'm not at all surprised. I've
    come to expect exactly this sort of ascenine behaviour from the people
    in Digital at the higher levels who are in charge of processes. Typical
    operating procedure is:
    
    	1, Let's authorize Fred to research X.
    	2, Fred researches X and comes back to report to the committee
    	3, Committee wants status quo, they wanted that from the beginning
    	   1) was only done to they could write a report to their higher
    	   ups saying they were doing something by authorizing something.
    	   Consequently committee now says how can we deep siz this
           initiative. Ha ha, no problem we can tell Fred to go and gather
    	   some quantitative information. We know damn well that in Digital
    	   this impossible. One of three things will happen:
    
    		a) Fred will get dispirited and we'll never hear from
    		   again. But mission accomplished we can now write a
    		  report saying:
    
    			"We investigated X, determined that no change is
    			 needed."
    
    		b) Fred will lose his funding or get reassigned to
                   something  else or will get TFSOed. Same result as a).
    
    		c) Fred will come and bug us again and we can label him
    		   an outcast or a trouble maker.
    
    I thought the supply chain initiatives were meant to fix this CRAP.
    
    I will now give a concrete axample of how we tried to make things easy
    for sales and how PAC (another useless committee) frustrated the
    product manager for one of my products.
    
    PM went to PAC with the proposal for DEC SNA 3270 Application Services
    back last May. We were proposing cluster tier type licensing. That was
    the defacto standard of the day.
    
    In June we concluded a deal with a third Party (INTEGRATA) and went to
    PAC to get approval for "MEP VT for DEC SNA 3270 Services". By this
    time the directive had come down not tp do cluster licenses anymore but
    to only do traditional part numbers (ie position 7 not position 9
    varies) so MEP VT was proposed like this. Now MEP VT has a prerequisite
    of DEC SNA 3270 Application Services. So we thought it made sense to
    ask to change 3270 Application Services to traditional license as well.
    
    The reply the PM got from PAC was "NO, we're not alowing people to
    change the licensing mechanism for existing products". We told them
    that this product hadn't FCSed yet so no licenses had been sold. Still
    "NO, it's policy. If we make an exception for you everybody else will
    want to change their license types".
    
    We gave up (too many other things to do) so now our two products have
    different license mechanisms. One traditional one cluster wide.
    But PAC is happy; bunch of bozos.
    
    Now if you'd like to hear the story of how I have been totally
    frustrated in finding SOMEONE who knew how to query the USCIC order
    system for a complete list of orders and quotes for a set of part
    numbers... Still haven't got it. I'll shut up now.
    
    To .0 Good luck. If you'd like some ammunition I suggest talking to  a
    few product managers who find the whole licensing/part number system as
    difficult to deal with from the top end as our field find it from the
    bottom end.
    
    One more suggestion. Write up your experience and sent it to Adriana
    Stecker (think that's the right name). She's Bob Palmer's right hand
    woman responsible for driving the Supply Chain change process.
    
    Dave

    RE: .26
    
    I suggest you do the following to make this happen...
    
    1)  Publish the names of everyone on the Business Practices Committee. 
    I will send them feedback directly.  In addition to that, have each of
    them publish their objectives and how they intend to meet them.  Also,
    I would like to see them do a cost justification of the current system
    and their participation in it.  We could have chaos for free.  What we
    have is chaos that costs us money.
    
    2)  Have all of them spend 3 days on the phones at 1-800-DEC-SALE
    answering questions about software.
    
    3)  Have all of them go to a field office (I will volunteer mine) and
    walk around asking for feedback.  Additionally they should actually
    solve a customer problem from start to finish.
    
    4)  Have them report to Ian Waring who actually has done what they are
    trying to do...he is only 2500 miles away.
    
    Until they feel the pain they cause first hand (not through an
    emisary) they will have no idea what is going on and no idea of how to
    fix it.  If the tone of this seems down, it is.  A prior note discussed
    the "virtues" of committees better than I can.  
    
    Ken Slattery

    The current system is unacceptable.  If the Committee doesn't agree
    on this fundamental point, then there's no way you'll ever be able
    to come up with "numbers" to justify any major changes.  You will
    forever be chasing shadows.
    
    I liked the previous reply about setting some standards and measuring
    the current system against it.  Then let the vested interests try and
    come up with "fixes" to the current system, while someone else does a
    straighforward re-engineering estimate.  May the best proposal win, as
    long as the new standards are met.
    
    Geoff Unland in Austin

    A previous reply hit the nail square and true, If you have not actually
    accomplished the task you are managing/developing/changing you do not
    have the qualifications necessary for the job.
    
    "If you have no idea what I do, don't tell me how to do it."
    
    .mike.

    Give them an exercise.  Below is a partial page from the price book. 
    VAX C is a pretty straight-forward product, so I did not pick a
    difficult one.  Have them find the part numbers for three different 
    configurations; pick any three, I don't care.  Have them keep track
    of the amount of time it took them.  Then have them write down a single
    part number followed by number of users; of course they will have to
    keep track of the amount of time required to do this.  Take the
    difference between those two numbers, multiply it by the number product
    licenses and upgrade we sell in a given year, and you will have an
    estimate of the cost savings.  Of course, I didn't get into the real
    `fun' part of media and documentation updates.
    
License                                Upgrade License
Rate                   from license rating in far left column to rating below

--------------------------------------------------------------------------------
  10| QL-015A9-JB |   20 | QL-015AQ-UC |  600 | QL-015AQ-UJ | 4800 | QL-015AQ-US
    | QT-015A9-LB |   50 | QL-015AQ-UD |  900 | QL-015AQ-UK | 6000 | QL-015AQ-UT
    | QL-015A9-SB |  100 | QL-015AQ-UE | 1200 | QL-015AQ-UL | 7200 | QL-015AQ-UU
    |             |  200 | QL-015AQ-UF | 1800 | QL-015AQ-UM | 9000 | QL-015AQ-UV
    |             |  300 | QL-015AQ-UG | 2400 | QL-015AQ-UN |      |
    |             |  400 | QL-015AQ-UH | 3600 | QL-015AQ-UQ |      |
--------------------------------------------------------------------------------
  20| QL-015A9-JC |   50 | QL-015AP-UD |  600 | QL-015AP-UJ | 3600 | QL-015AP-UQ
    | QT-015A9-LC |  100 | QL-015AP-UE |  900 | QL-015AP-UK | 4800 | QL-015AP-US
    | QL-015A9-SC |  200 | QL-015AP-UF | 1200 | QL-015AP-UL | 6000 | QL-015AP-UT
    |             |  300 | QL-015AP-UG | 1800 | QL-015AP-UM | 7200 | QL-015AP-UU
    |             |  400 | QL-015AP-UH | 2400 | QL-015AP-UN | 9000 | QL-015AP-UV
--------------------------------------------------------------------------------
  50| QL-015A9-JD |  100 | QL-015AN-UE |  900 | QL-015AN-UK | 4800 | QL-015AN-US
    | QT-015A9-LD |  200 | QL-015AN-UF | 1200 | QL-015AN-UL | 6000 | QL-015AN-UT
    | QL-015A9-SD |  300 | QL-015AN-UG | 1800 | QL-015AN-UM | 7200 | QL-015AN-UU
    |             |  400 | QL-015AN-UH | 2400 | QL-015AN-UN | 9000 | QL-015AN-UV
    |             |  600 | QL-015AN-UJ | 3600 | QL-015AN-UQ |      |
--------------------------------------------------------------------------------
 100| QL-015A9-JE |  200 | QL-015AJ-UF | 1200 | QL-015AJ-UL | 6000 | QL-015AJ-UT
    | QT-015A9-LE |  300 | QL-015AJ-UG | 1800 | QL-015AJ-UM | 7200 | QL-015AJ-UU
    | QL-015A9-SE |  400 | QL-015AJ-UH | 2400 | QL-015AJ-UN | 9000 | QL-015AJ-UV
    |             |  600 | QL-015AJ-UJ | 3600 | QL-015AJ-UQ |      |
    |             |  900 | QL-015AJ-UK | 4800 | QL-015AJ-US |      |
--------------------------------------------------------------------------------
 200| QL-015A9-JF |  300 | QL-015AH-UG | 1200 | QL-015AH-UL | 4800 | QL-015AH-US
    | QT-015A9-LF |  400 | QL-015AH-UH | 1800 | QL-015AH-UM | 6000 | QL-015AH-UT
    | QL-015A9-SF |  600 | QL-015AH-UJ | 2400 | QL-015AH-UN | 7200 | QL-015AH-UU
    |             |  900 | QL-015AH-UK | 3600 | QL-015AH-UQ | 9000 | QL-015AH-UV
--------------------------------------------------------------------------------
 300| QL-015A9-JG |  400 | QL-015AB-UH | 1800 | QL-015AB-UM | 6000 | QL-015AB-UT
    | QT-015A9-LG |  600 | QL-015AB-UJ | 2400 | QL-015AB-UN | 7200 | QL-015AB-UU
    | QL-015A9-SG |  900 | QL-015AB-UK | 3600 | QL-015AB-UQ | 9000 | QL-015AB-UV
    |             | 1200 | QL-015AB-UL | 4800 | QL-015AB-US |      |
--------------------------------------------------------------------------------
 400| QL-015A9-JH |  600 | QL-015AG-UJ | 2400 | QL-015AG-UN | 7200 | QL-015AG-UU
    | QT-015A9-LH |  900 | QL-015AG-UK | 3600 | QL-015AG-UQ | 9000 | QL-015AG-UV
    | QL-015A9-SH | 1200 | QL-015AG-UL | 4800 | QL-015AG-US |      |
    |             | 1800 | QL-015AG-UM | 6000 | QL-015AG-UT |      |
--------------------------------------------------------------------------------
 600| QL-015A9-JJ |  900 | QL-015AE-UK | 2400 | QL-015AE-UN | 6000 | QL-015AE-UT
    | QT-015A9-LJ | 1200 | QL-015AE-UL | 3600 | QL-015AE-UQ | 7200 | QL-015AE-UU
    | QL-015A9-SJ | 1800 | QL-015AE-UM | 4800 | QL-015AE-US | 9000 | QL-015AE-UV
--------------------------------------------------------------------------------
 900| QL-015A9-JK | 1200 | QL-015A4-UL | 3600 | QL-015A4-UQ | 7200 | QL-015A4-UU
    | QT-015A9-LK | 1800 | QL-015A4-UM | 4800 | QL-015A4-US | 9000 | QL-015A4-UV
    | QL-015A9-SK | 2400 | QL-015A4-UN | 6000 | QL-015A4-UT |      |
--------------------------------------------------------------------------------
1200| QL-015A9-JL | 1800 | QL-015AD-UM | 4800 | QL-015AD-US | 9000 | QL-015AD-UV
    | QT-015A9-LL | 2400 | QL-015AD-UN | 6000 | QL-015AD-UT |      |
    | QL-015A9-SL | 3600 | QL-015AD-UQ | 7200 | QL-015AD-UU |      |
--------------------------------------------------------------------------------
1800| QL-015A9-JM | 2400 | QL-015AU-UN | 6000 | QL-015AU-UT |      |
    | QT-015A9-LM | 3600 | QL-015AU-UQ | 7200 | QL-015AU-UU |      |
    | QL-015A9-SM | 4800 | QL-015AU-US | 9000 | QL-015AU-UV |      |
--------------------------------------------------------------------------------
2400| QL-015A9-JN | 3600 | QL-015AV-UQ | 7200 | QL-015AV-UU |      |
    | QT-015A9-LN | 4800 | QL-015AV-US | 9000 | QL-015AV-UV |      |
    | QL-015A9-SN | 6000 | QL-015AV-UT |      |             |      |
--------------------------------------------------------------------------------
3600| QL-015A9-JQ | 4800 | QL-015A6-US | 9000 | QL-015A6-UV |      |
    | QT-015A9-LQ | 6000 | QL-015A6-UT |      |             |      |
    | QL-015A9-SQ | 7200 | QL-015A6-UU |      |             |      |
--------------------------------------------------------------------------------
4800| QL-015A9-JS | 6000 | QL-015A5-UT |      |             |      |
    | QT-015A9-LS | 7200 | QL-015A5-UU |      |             |      |
    | QL-015A9-SS | 9000 | QL-015A5-UV |      |             |      |
--------------------------------------------------------------------------------
6000| QL-015A9-JT | 7200 | QL-015AT-UU |      |             |      |
    | QT-015A9-LT | 9000 | QL-015AT-UV |      |             |      |
    | QL-015A9-ST |      |             |      |             |      |
--------------------------------------------------------------------------------
7200| QL-015A9-JU | 9000 | QL-015A7-UV |      |             |      |
    | QT-015A9-LU |      |             |      |             |      |
    | QL-015A9-SU |      |             |      |             |      |
--------------------------------------------------------------------------------
9000| QL-015A9-JV |      |             |      |             |      |
    | QT-015A9-LV |      |             |      |             |      |
    | QL-015A9-SV |      |             |      |             |      |
--------------------------------------------------------------------------------

    RE: 37
    
    I sincerely hope you didn't spend too much time typing in all those
    part #s for VAX C.......'cause the cluster-wide upgrade part #s you
    used in the example are no longer saleable :-)
    
    But then, if you had called 1-800-DEC-SALE first, we could have told
    you that :-)
    
    Karen
    
    PS:  You must be using an out-dated price book (another problem for
    	 field people, obtaining current price books, that is); the Q1
         U.S. Software Price Book has removed all such part numbers.
    

    Most of the comments entered so far, I hear on a daily basis; none
    of the notes are exaggerated.  What is so crazy is that we are
    carrying this format into the PC space!!
    
    A PC based product (send me mail if you want the name....or if you've
    tried to configure it you'll probably guess the name); has 19 part #s!!
    Granted, the 19 part numbers are not just the licenses, this includes
    media & doc kits and SPS service part #s.
    
    A few weeks ago I had been working with a sales rep trying to make
    sure he had all the pieces for his quote.  After we have spent approx-
    imately 40 minutes on the phone, the rep said very politely "Karen, 
    you've been a great help, but I've just decided I'm not going to pur-
    sue this any further.  I spent 2 afternoons before I called you, we've
    been on the phone for over 1/2 hour; my customer doesn't have a 
    burning desire for this product so I'm not going to waste any more time
    on it.  My customer would probably like the product, but DEC probably
    has LOST money on it just in the time spent trying to quote the d**n
    thing!!"
    
    This really drove home Mr. Palmer's statement that DEC cannot continue
    to be all things to all companies.  If the E/U were dealing with any
    company other than DEC, the E/U wouldn't give a second thought to
    buying a new version if/when it is released - remember this is a prod-
    duct targeted to bring PCs into a LAN.  What makes this scenario
    sadder, is that the product was presented as a "package" to eliminate
    confusion in selling......
    
    Now some people might fault that sales rep; I don't.  He came to the
    only conclusion possible - unless a customer is holding a gun to his
    head and perhaps a larger sale is at stake, why waste all the time as
    he had done?
    
    The product itself seems to be a rather good one; if it fails to make
    its mark it will be because it is too complex to configure :-(
    
    Karen