Market Share v. User Base - Linux

This is a discussion on Market Share v. User Base - Linux ; I note that figures for "Market Share" are frequently posted in this group. One frequent source seems to be Net Applications:- http://marketshare.hitslink.com/report.aspx?qprid=9 Using those figures, I find that over a period of 10 months to April 2008, the changes that ...

+ Reply to Thread
Results 1 to 19 of 19

Thread: Market Share v. User Base

  1. Market Share v. User Base

    I note that figures for "Market Share" are frequently posted in this
    group.
    One frequent source seems to be Net Applications:-
    http://marketshare.hitslink.com/report.aspx?qprid=9

    Using those figures, I find that over a period of 10 months to April
    2008, the changes that have occurred are:-

    Microsoft: From 93.34% to 91.56% (-1.78%)
    Mac: From 6.03% to 7.37%
    Linux From 0.43% to 0.72%

    Translated into "per annum" figures (12 months, not 10), this gives us:-

    Microsoft -2.14%
    Mac +1.61%
    Linux +0.35%

    OK, so MS is vast, and can *afford* to lose 2% of market share?
    So what's my point?
    My point is that those figures are *not* market shares!
    They are the User Base shares!

    What's the difference?
    Well, quite large!

    Change these guesses of mine how you like, but the over-riding principle
    remains the same:-

    - Assume that a year ago, there were 1,100M Internet-connected
    computers.
    - Assume that 100M of those were scrapped over the year, leaving 1,000M
    - Assume that 200M new machines were sold over the year.

    *Question* :-
    What percentage of those 200M new computers would need to use which OS
    to give the final figures (bearing in mind that each needs to sell
    its "old" percentage just to stay where it is)?

    *Answer* :
    Microsoft 82.66% (not 91.56)
    Mac 14.07% (not 7.37)
    Linux 2.17% (not 0.72)

    Interesting?
    .... or perhaps my sums are wrong? :-)


  2. Re: Market Share v. User Base

    * bbgruff peremptorily fired off this memo:

    > *Question* :-
    > What percentage of those 200M new computers would need to use which OS
    > to give the final figures (bearing in mind that each needs to sell
    > its "old" percentage just to stay where it is)?
    >
    > *Answer* :
    > Microsoft 82.66% (not 91.56)
    > Mac 14.07% (not 7.37)
    > Linux 2.17% (not 0.72)
    >
    > Interesting?
    > ... or perhaps my sums are wrong? :-)


    The calculus of finite differences will set you free.

    --
    As we look ahead into the next century, leaders will be those who empower
    others.
    -- Bill Gates

  3. Re: Market Share v. User Base


    "bbgruff" wrote in message
    news:68e9tuF2rlnqnU1@mid.individual.net...

    > ... or perhaps my sums are wrong? :-)
    >


    Not just your sums, your whole method of approach. Market shares are
    measured directly, by getting the total value of product shipments and
    dividing into each contributor's revenue total. When you do this, you find
    that Microsoft has some 99+% of the actual market for their product on the
    desktop. Apple does not sell MacOS except to upgrade Mac computers. Very
    little Linux is sold for the desktop and it hardly registers. Things that
    you cannot displace with your own sales are simply not a part of your
    market. It may feel more correct to think that Apple is competing with
    Microsoft, but they are really competing with Dell and HP. Microsoft can't
    weep over a Macintosh sale any more than they can care about someone buying
    a TV set and Wii rather than buying a new computer. The products are not
    competitive.

    A to your sums, the correct method would be to first see if the data were
    consistent and then calculate the expected standard deviation of the data to
    see if the variations were natural or if they represented a statistically
    significant change. Also, was there any change in the base, as you
    conclude, or was there just a change in usage patterns. Since you are using
    such an indirect method, you have o consider the repeatibility and
    reliability of the measurements. None of that is evident.


  4. Re: Market Share v. User Base


    >
    > Interesting?
    > ... or perhaps my sums are wrong? :-)


    The effect you are describing is real, and it probably explains why
    Mac sales are up 50% over last year but their market share is up only
    roughly 20% (relatively speaking). In the case of Linux, it may be
    more complicated, for example, some of those 100 machines that were
    thrown away might actually have been converted to Linux.

  5. Re: Market Share v. User Base

    nessuno@wigner.berkeley.edu wrote:

    >>
    >> Interesting?
    >> ... or perhaps my sums are wrong? :-)

    >
    > The effect you are describing is real, and it probably explains why
    > Mac sales are up 50% over last year but their market share is up only
    > roughly 20% (relatively speaking). In the case of Linux, it may be
    > more complicated, for example, some of those 100 machines that were
    > thrown away might actually have been converted to Linux.


    Yep.
    A lot of the figures that I see (particularly Mac v. Windows) make sense
    once you twig to the principle, don't they?
    The absolute values will vary, of course, depending on where you get the
    figures. However, one consistent trend that I do believe I've seen is
    that both Mac and Linux have gained considerably over the past year, at
    the expense of Windows.
    From memory, I thought I'd seen a figure of 140M copies of Vista in its
    first 14 months. The announcement was along the lines of "so very
    many". My own reaction was "How come so few?"

    Really, I'm not at all sure what one should do about "conversions to
    Linux" as far as a "market share" concept is concerned. For instance,
    I bought this very machine with Vista on it, and it's now theoretically
    a dual-boot machine, Vista/Ubuntu. That was a "sale" for MS, so counts
    to their market share. However, since I booted into Vista on it only
    during the first week, and haven't done so since last November(!), it
    will be contributing only to the Linux "user base" as seen by 'net
    statistics.
    In theory then, MS could have 100% of the market, every machine sold
    anywhere could be sold with Windows on it, and yet (theoretically)
    everybody could be using Linux on them! In that case, we'd have 100%
    MS market, 100% Linux use!

    I think I'll stay with my own concept of the situation, and judge market
    share from the impact of the various contenders on the user base:-)


  6. Re: Market Share v. User Base

    bbgruff wrote:

    > My point is that those figures are *not* market shares!
    > They are the User Base shares!


    > Interesting?
    > ... or perhaps my sums are wrong? :-)


    Your thinking is right.

    (year-on-year change in usage share)
    x
    (average life of a computer in years)
    =
    (market share)

  7. Re: Market Share v. User Base

    bbgruff wrote:
    > nessuno@wigner.berkeley.edu wrote:
    >
    > From memory, I thought I'd seen a figure of 140M copies of Vista in its
    > first 14 months. The announcement was along the lines of "so very
    > many". My own reaction was "How come so few?"


    Yes. Considering there were 270 million PCs sold in 2007, Vista was
    going onto fewer than half of new computers during those 14 months.

    http://www.gartner.com/it/page.jsp?id=584210

  8. Re: Market Share v. User Base

    On 2008-05-07, amicus_curious wrote:
    >
    > "bbgruff" wrote in message
    > news:68e9tuF2rlnqnU1@mid.individual.net...
    >
    >> ... or perhaps my sums are wrong? :-)
    >>

    >
    > Not just your sums, your whole method of approach. Market shares are
    > measured directly, by getting the total value of product shipments and
    > dividing into each contributor's revenue total. When you do this, you find


    ....which is ultimately meaningless.

    This isn't about dollars, it's about bodies and boxes.

    > that Microsoft has some 99+% of the actual market for their product on the
    > desktop. Apple does not sell MacOS except to upgrade Mac computers. Very
    > little Linux is sold for the desktop and it hardly registers. Things that
    > you cannot displace with your own sales are simply not a part of your
    > market. It may feel more correct to think that Apple is competing with
    > Microsoft, but they are really competing with Dell and HP. Microsoft can't
    > weep over a Macintosh sale any more than they can care about someone buying
    > a TV set and Wii rather than buying a new computer. The products are not
    > competitive.


    No.

    First Apple must compete against Microsoft.

    It is the Microsoft "ecosystem" that sells HPs or Dells. Without this,
    the major name brand PCs don't have any thing to offer. Even then, they
    are all generic and pretty much completely interchangeable. For the
    purposes of this discussion they can be all lumped in together.

    When Dell loses, Microsoft loses.

    When Dell ships a Linux server box, Microsoft loses too.

    Microsoft might still lose in the long run even when Dell ships a
    box with Windows on it. It's hard to measure that.

    >
    > A to your sums, the correct method would be to first see if the data were
    > consistent and then calculate the expected standard deviation of the data to
    > see if the variations were natural or if they represented a statistically
    > significant change. Also, was there any change in the base, as you
    > conclude, or was there just a change in usage patterns. Since you are using
    > such an indirect method, you have o consider the repeatibility and
    > reliability of the measurements. None of that is evident.
    >



    --

    iTunes is not progressive. It's a throwback. |||
    / | \

    Posted Via Usenet.com Premium Usenet Newsgroup Services
    ----------------------------------------------------------
    http://www.usenet.com

  9. Re: Market Share v. User Base

    On 2008-05-07, bbgruff claimed:

    > A lot of the figures that I see (particularly Mac v. Windows) make sense
    > once you twig to the principle, don't they?
    > The absolute values will vary, of course, depending on where you get the
    > figures. However, one consistent trend that I do believe I've seen is
    > that both Mac and Linux have gained considerably over the past year, at
    > the expense of Windows.
    > From memory, I thought I'd seen a figure of 140M copies of Vista in its
    > first 14 months. The announcement was along the lines of "so very
    > many". My own reaction was "How come so few?"


    100M are likely still sitting someplace, waiting for buyers.

    > Really, I'm not at all sure what one should do about "conversions to
    > Linux" as far as a "market share" concept is concerned. For instance,
    > I bought this very machine with Vista on it, and it's now theoretically
    > a dual-boot machine, Vista/Ubuntu. That was a "sale" for MS, so counts
    > to their market share. However, since I booted into Vista on it only
    > during the first week, and haven't done so since last November(!), it
    > will be contributing only to the Linux "user base" as seen by 'net
    > statistics.


    It's worse than that.

    MS sells Vista to an OEM. The OEM sells Vista to the consumer. The
    consumer demands XP. The OEM sells XP to the consumer. MS gets 2 sales.

    By that count, that's 2 machines with Windross on them. The reality is
    only half that many.

    Such creative calculating allows the numbers of users of proprietary
    crapware appear to be much larger than they are. Not double, because
    not everyone demands XP right off (some pretend they're happy with
    Vista for a couple of months before they can't even personally bear to
    tell that lie any more). And many don't buy new machines right off. But
    the numbers of sales appear to equate to the numbers of total users,
    and it's inaccurate.

    > In theory then, MS could have 100% of the market, every machine sold
    > anywhere could be sold with Windows on it, and yet (theoretically)
    > everybody could be using Linux on them! In that case, we'd have 100%
    > MS market, 100% Linux use!


    It would be interesting to see 100% of machines sold in a year have
    only Vista on them, and then compare that to the number of total copies
    of Winders sold in the same period (not using MS' numbers, but numbers
    from actual retailers of the products). Preferably only counting the XP
    copies sold to the same people who were sold machines with Vista on
    them. I wouldn't be surprised to find MS has 130% market share based on
    the total copies sold.

    The monopoly is probably retarded enough to make that claim as a
    promotion of how healthy they are, too. The lapdog would repeat it
    endlessly as fact as well.

    --
    Windows doesn't have any bugs. It just develops random features.

  10. Re: Market Share v. User Base


    "Matt" wrote in message
    news:0NpUj.2673$Cn4.236@news02.roc.ny...
    > bbgruff wrote:
    >
    >> My point is that those figures are *not* market shares!
    >> They are the User Base shares!

    >
    >> Interesting?
    >> ... or perhaps my sums are wrong? :-)

    >
    > Your thinking is right.
    >
    > (year-on-year change in usage share)
    > x
    > (average life of a computer in years)
    > =
    > (market share)


    That would gain you an F at most business schools, but that is not so
    important. "Market share" remains as the share of the revenues of products
    sold in the market regardless. It is a metric that matters to those who
    sell in the market and no one else other than as an intellectual curiosity.
    You folk seem to want to structure some kind of number that tells the score
    and, more importantly, bolsters your forlorn hope that Microsoft is on the
    decline. Good luck!


  11. Re: Market Share v. User Base

    "amicus_curious" writes:

    > "Matt" wrote in message
    > news:0NpUj.2673$Cn4.236@news02.roc.ny...
    >> bbgruff wrote:
    >>
    >>> My point is that those figures are *not* market shares!
    >>> They are the User Base shares!

    >>
    >>> Interesting?
    >>> ... or perhaps my sums are wrong? :-)


    Only if you are an idiot and purposely obfuscate the meaning. Its
    blatantly clear to *ANYONE* what market share means in terms of
    distributing a desktop OS - for free or not.

    >>
    >> Your thinking is right.
    >>


    No it isn't.

    >> (year-on-year change in usage share)
    >> x
    >> (average life of a computer in years)
    >> =
    >> (market share)

    >
    > That would gain you an F at most business schools, but that is not so
    > important. "Market share" remains as the share of the revenues of
    > products sold in the market regardless. It is a metric that matters
    > to those who sell in the market and no one else other than as an
    > intellectual curiosity. You folk seem to want to structure some kind
    > of number that tells the score and, more importantly, bolsters your
    > forlorn hope that Microsoft is on the decline. Good luck!
    >


    Only in COLA could a term be messed with so much to support their crazy
    numbers. Market Share/ User % is the same thing (obviously) when talking
    about a free OS v a purchased one. And the Linux usage % languishes
    below 1% IMO - and mainly due to cess pits like COLA.

    I showed my Lenny laptop last night and it blew them away in terms of
    how well it ran (its an old X30=. I obviously didnt power up Open
    Office, but showed some other stuff running like various modes in emacs,
    opengl screen saver and other eye candy. The wandering mode in the
    wireless settings now work flawlessly and since the last update
    "hibernate" works 95% of the time too when I close the laptop lid. This
    is REAL advocacy - not blathering on here about how Linux has (guffaw)
    17% of the "market". Even my dog knows thats not true.

  12. Re: Market Share v. User Base

    Matt wrote:
    > bbgruff wrote:
    >
    >> My point is that those figures are *not* market shares!
    >> They are the User Base shares!

    >
    >> Interesting?
    >> ... or perhaps my sums are wrong? :-)

    >
    > Your thinking is right.
    >
    > (year-on-year change in usage share)
    > x
    > (average life of a computer in years)
    > =
    > (market share)



    Sorry, this equation is wrong. I should have written:

    (year-on-year change in usage share)
    x
    (average life of a computer in years)
    =
    (market share - usage share)

  13. Re: Market Share v. User Base

    Matt wrote:
    > Matt wrote:
    >> bbgruff wrote:
    >>
    >>> My point is that those figures are *not* market shares!
    >>> They are the User Base shares!


    > (year-on-year change in usage share)
    > x
    > (average life of a computer in years)
    > =
    > (market share - usage share)


    Let
    us = usage share
    ms = market share
    T = average lifespan of a computer
    t = time

    Then

    (d/dt)(us) = (ms - us)/T

    and

    the change in usage share over a time interval [t1, t2],

    us(t2) - us(t1) = (ms - us) * ((t2 - t1))/T

  14. Re: Market Share v. User Base

    amicus_curious wrote:

    > From a business standpoint, what you are proposing is immaterial


    Then please feel free to ignore my posts.

  15. Re: Market Share v. User Base

    amicus_curious wrote:
    >
    > "Matt" wrote in message
    > news:0NpUj.2673$Cn4.236@news02.roc.ny...
    >> bbgruff wrote:
    >>
    >>> My point is that those figures are *not* market shares!
    >>> They are the User Base shares!

    >>
    >>> Interesting?
    >>> ... or perhaps my sums are wrong? :-)

    >>
    >> Your thinking is right.
    >>
    >> (year-on-year change in usage share)
    >> x
    >> (average life of a computer in years)
    >> =
    >> (market share)

    >
    > That would gain you an F at most business schools,


    If you mean to say that that equation has an error, you are correct, and
    elsewhere on the thread I have corrected the error.

  16. Re: Market Share v. User Base

    nessuno@wigner.berkeley.edu wrote:
    >> Interesting?
    >> ... or perhaps my sums are wrong? :-)

    >
    > The effect you are describing is real, and it probably explains why
    > Mac sales are up 50% over last year but their market share is up only
    > roughly 20% (relatively speaking).



    How about trying:

    (d/dt)(usage share)
    =
    (market share - usage share) / (average lifespan of a computer)

    using the definitions at

    http://en.wikipedia.org/wiki/Market_share
    http://en.wikipedia.org/wiki/Usage_s...rating_systems

    on your Mac data?

  17. Re: Market Share v. User Base

    bbgruff wrote:
    > I note that figures for "Market Share" are frequently posted in this
    > group.
    > One frequent source seems to be Net Applications:-
    > http://marketshare.hitslink.com/report.aspx?qprid=9
    >
    > Using those figures, I find that over a period of 10 months to April
    > 2008, the changes that have occurred are:-
    >
    > Microsoft: From 93.34% to 91.56% (-1.78%)
    > Mac: From 6.03% to 7.37%
    > Linux From 0.43% to 0.72%
    >
    > Translated into "per annum" figures (12 months, not 10), this gives us:-
    >
    > Microsoft -2.14%
    > Mac +1.61%
    > Linux +0.35%


    > My point is that those figures are *not* market shares!
    > They are the User Base shares!


    > - Assume that a year ago, there were 1,100M Internet-connected
    > computers.
    > - Assume that 100M of those were scrapped over the year, leaving 1,000M
    > - Assume that 200M new machines were sold over the year.
    >
    > *Question* :-
    > What percentage of those 200M new computers would need to use which OS
    > to give the final figures (bearing in mind that each needs to sell
    > its "old" percentage just to stay where it is)?
    >
    > *Answer* :
    > Microsoft 82.66% (not 91.56)
    > Mac 14.07% (not 7.37)
    > Linux 2.17% (not 0.72)



    The number of Windows computers a year ago minus the number of Windows
    computers scrapped during the year plus the number of new Windows
    computers equals the present number of Windows computers:

    (0.9344)(1000M) - (0.9344)(100M) + (w)(200M) = (0.9156)(1100M)
    w = (1007.16M - 934.4M + 93.4M) / 200M
    w = 166.2M / 200M
    w = 0.831
    w = 83.1%

    I used averaged values for the year. Maybe you divided the year into
    smaller periods and made some assumptions of linearity.

    It would help to know usage shares from five years ago, since a higher
    fraction of computers were running Windows then. That means the
    computers being scrapped now are more likely than average to have been
    running Windows.

    How about showing your calculations?

  18. Re: Market Share v. User Base

    Matt writes:

    > bbgruff wrote:
    >> I note that figures for "Market Share" are frequently posted in this
    >> group.
    >> One frequent source seems to be Net Applications:-
    >> http://marketshare.hitslink.com/report.aspx?qprid=9
    >>
    >> Using those figures, I find that over a period of 10 months to April
    >> 2008, the changes that have occurred are:-
    >>
    >> Microsoft: From 93.34% to 91.56% (-1.78%)
    >> Mac: From 6.03% to 7.37%
    >> Linux From 0.43% to 0.72%
    >>
    >> Translated into "per annum" figures (12 months, not 10), this gives us:-
    >>
    >> Microsoft -2.14%
    >> Mac +1.61%
    >> Linux +0.35%

    >
    >> My point is that those figures are *not* market shares!
    >> They are the User Base shares!

    >
    >> - Assume that a year ago, there were 1,100M Internet-connected
    >> computers.
    >> - Assume that 100M of those were scrapped over the year, leaving 1,000M
    >> - Assume that 200M new machines were sold over the year.
    >>
    >> *Question* :-
    >> What percentage of those 200M new computers would need to use which OS
    >> to give the final figures (bearing in mind that each needs to sell
    >> its "old" percentage just to stay where it is)?
    >>
    >> *Answer* :
    >> Microsoft 82.66% (not 91.56)
    >> Mac 14.07% (not 7.37)
    >> Linux 2.17% (not 0.72)

    >
    >
    > The number of Windows computers a year ago minus the number of Windows
    > computers scrapped during the year plus the number of new Windows
    > computers equals the present number of Windows computers:
    >
    > (0.9344)(1000M) - (0.9344)(100M) + (w)(200M) = (0.9156)(1100M)
    > w = (1007.16M - 934.4M + 93.4M) / 200M
    > w = 166.2M / 200M
    > w = 0.831
    > w = 83.1%
    >
    > I used averaged values for the year. Maybe you divided the year into
    > smaller periods and made some assumptions of linearity.
    >
    > It would help to know usage shares from five years ago, since a higher
    > fraction of computers were running Windows then. That means the
    > computers being scrapped now are more likely than average to have been
    > running Windows.
    >
    > How about showing your calculations?


    Well for a start we can rubbish your silly percentages since they do not
    consider any "scrapping" or uptake of any other systems!

  19. Re: Market Share v. User Base

    Matt wrote:
    > bbgruff wrote:
    >> I note that figures for "Market Share" are frequently posted in this
    >> group.
    >> One frequent source seems to be Net Applications:-
    >> http://marketshare.hitslink.com/report.aspx?qprid=9
    >>
    >> Using those figures, I find that over a period of 10 months to April
    >> 2008, the changes that have occurred are:-
    >>
    >> Microsoft: From 93.34% to 91.56% (-1.78%)
    >> Mac: From 6.03% to 7.37%
    >> Linux From 0.43% to 0.72%
    >>
    >> Translated into "per annum" figures (12 months, not 10), this gives us:-
    >>
    >> Microsoft -2.14%
    >> Mac +1.61%
    >> Linux +0.35%

    >
    >> My point is that those figures are *not* market shares!
    >> They are the User Base shares!

    >
    >> - Assume that a year ago, there were 1,100M Internet-connected
    >> computers.
    >> - Assume that 100M of those were scrapped over the year, leaving 1,000M
    >> - Assume that 200M new machines were sold over the year.
    >>
    >> *Question* :-
    >> What percentage of those 200M new computers would need to use which OS
    >> to give the final figures (bearing in mind that each needs to sell
    >> its "old" percentage just to stay where it is)?
    >>
    >> *Answer* :
    >> Microsoft 82.66% (not 91.56)
    >> Mac 14.07% (not 7.37)
    >> Linux 2.17% (not 0.72)

    >
    >
    > The number of Windows computers a year ago minus the number of Windows
    > computers scrapped during the year plus the number of new Windows
    > computers equals the present number of Windows computers:
    >
    > (0.9344)(1000M) - (0.9344)(100M) + (w)(200M) = (0.9156)(1100M)
    > w = (1007.16M - 934.4M + 93.4M) / 200M
    > w = 166.2M / 200M
    > w = 0.831
    > w = 83.1%



    We could notice that Windows usage on five-year-old computers is higher
    than on average-aged computers. Suppose that 97% of retiring computers
    are running Windows.

    Then reasoning as before:

    (0.9344)(1000M) - (0.97)(100M) + (w)(200M) = (0.9156)(1100M)
    w = (1007.2M - 934.4M + 97M) / 200M
    w = 169.8M / 200M
    w = 0.849
    w ~~= 85%

    These calculations illustrate a way of thinking about usage share vs.
    market share, and some of the numbers are known to be inaccurate. For
    instance, the number of computers sold in 2007 was actually 271.2M, not
    200M.

    http://www.gartner.com/it/page.jsp?id=584210

+ Reply to Thread