Debunking the 5% Myth

[JamesDPS]

I'm glad for people pointing out the distinction between market share and install base... I guess I hadn't thought that through -- clearly, Macs have a greater longevity than other computers (I'm still using my 4-year-old Powerbook and an - granted, upgraded - old PowerMac 9500, and I expect a new G5, when I get one, to last me a good 5 or 6 years); it kind of makes sense, now, how Apple can have a very low market share, but still have a decent install base that attracts developers. Of course, getting that install base to switch to OS X as much as possible is the key, and that was one of SJ's topics during his MWSF keynote (that is, how successful Apple has been with the changeover).

Anyway this whole thread reminds me of a book I read in high school

 

[cubist]

... 520,000 AS400's, 210,000 mainframes, 3 million Unix servers, 9 million
NT servers, 240 million Windows PC's, 32 million Macintosh PC's, and 3
million Linux PC's. ...

Installed base numbers are very difficult to obtain. These numbers are likely mostly guesses. Who knows how many PCs and Macs are sitting in landfills, in garages, in attics? I have 20 computers in my house; some are connected and working, some are in storage, some are awaiting being thrown out.

Market share numbers are much easier to obtain: how many units sold this quarter? That's why we see them being quoted more often.

I recall about 5 years ago the results of a fairly large survey of PC users, that over half of them were still using Windows 3.1. Microsoft has never been happy about the upgrade rates. Most of the people I know with PCs at home are still using Windows 98. There isn't a single thing that XP does, that 98 doesn't do, that interests them in any way.

By contrast, Mac OS X is such a huge leap beyond what we used to do with 9, I wouldn't even consider going back. And it is a huge leap beyond Windows. Goodbye and good riddance, Windows, with your proprietary APIs and your DLLs and your _far _pascal _dllexport STDCALL MakeProcInstance() gibberish.

 

[Westside guy]

Repeated sampling will show you that the representation stands for the sample group.

Yes, and as I stated in my first post, the sample group likely is pretty much the whole group here; unless you believe there's the computer user equivalent of "dark matter" out there.

Generalizability is extremely debatable, and always has been, especially in something as socioeconomically stratified as computer usage.

Of course, we don't have anything to base this on aside from a very loosely defined naturalistic observation, so it's even harder to apply to a broader group. The confounds that immediately leap to mind are factors like the cost of macs (implying certain economic trends in users), the demographics of computer use in general, the draw of certain sites, and software errors in either browser or site statistic programs.

We're not discussing why Macs have the market share they do. The argument was that the statistics Google gathers on their visitors do not reflect the group of computer users as a whole. Unless you have evidence that there is some correlation between socioeconomic status and Google usage, this seems pretty much irrelevant.

There is no such thing as a perfect statistical method.

Yup. No argument here. But that hasn't ended this thread...

I use a proxy that screens me out of Google's information collections system. I can't tell you if it reports me as a mac, Windows, or Linux user, since I don't know what the proxy is running on. That's just one possible explanation for the difference in statistical numbers, and that leaves out things like competing search engines.

The first part is hard to argue with without doing some research on how proxies handle things like Google - do they proxy a Google search at all? I suspect, though, that it's not a significant factor since most of the proxy servers I'm personally aware of are running on Linux, and Google's Zeitgeist reports Linux usage at 1%.

The good site statistics programs have ways of attempting to ignore proxys completely, by checking things such as session length - I don't really know how effective that is though.

 

[thatwendigo]

Yes, and as I stated in my first post, the sample group likely is pretty much the whole group here; unless you believe there's the computer user equivalent of "dark matter" out there.

Actually, I know I'm a bad example, but I happen know that my family is in possessiong of at least four macs that are in running order, capable of runnings OS X (and I believe are), but not online. This is largely for reasons of switchings and NAT adressing off of thee and four-port broadband routers, which the manufacturers haven't made nearly as easy as Apple made their interface. I've got it working for my setup, but haven't fooled around with the other hardware in other domiciles.

Also, I do deny that there's any evidence that the sample group is representative. It's up to you to prove that, since you're making a positive knowledge claim and I've pointed out some problems with the basic assumption.

We're not discussing why Macs have the market share they do. The argument was that the statistics Google gathers on their visitors do not reflect the group of computer users as a whole. Unless you have evidence that there is some correlation between socioeconomic status and Google usage, this seems pretty much irrelevant.

They reflect the group of computer users who access Google. Period.

The first part is hard to argue with without doing some research on how proxies handle things like Google - do they proxy a Google search at all? I suspect, though, that it's not a significant factor since most of the proxy servers I'm personally aware of are running on Linux, and Google's Zeitgeist reports Linux usage at 1%.

The good site statistics programs have ways of attempting to ignore proxys completely, by checking things such as session length - I don't really know how effective that is though.

My understanding of the proxy I use is that it strips my IP from the search (substituting theirs) and bars the cookie from setting, which prevents Google from tagging my search results within their system. The proxy maintains system access logs for legal reasons, but they're deleted every seven days and don't contain the kind of information that Google harvests.

 

[Westside guy]

Also, I do deny that there's any evidence that the sample group is representative. It's up to you to prove that, since you're making a positive knowledge claim and I've pointed out some problems with the basic assumption.

This is descending into a case of "you say this, but I say that". I haven't seen you point out anything that doesn't seem more than tangential at best, but it seems obvious neither of us is going to convice the other.

 

[applebum]

In the last 10 months about 3% of my site's visitors were on Macs ...

If a site is not at all tangentially related to the subject being measured (for example, mine's a gardening site), it's hard to argue that - assuming you get enough counts to be statistically relevant - the site's visit numbers are somehow being skewed by the topic.

So, what you are saying is that 3% of the people with computers that are online, who are interested in gardening, are using Macs.

I would also guess that 100% of the people that visited your site have some knowledge of computers, would you say those numbers are statistically accurate for all people?

Your numbers may match up with the actual marketshare of Apple, but I tend to think it may be more coincidence than anything else. The only way I could see that site numbers were relevant was if every non-computer oriented website reported the exact same numbers.

 

[Zaty]

Guys, there's one thing I don't understand at all: As a recent switcher I wonder why Apple's market share is such an important matter. Does it really matter if Apple's market share is 2% or 10%? Does it make your Mac better or worse? I don't think so. As long as we don't have to worry if Apple may go out of business, their market share or installed baser for that matter is mere statistics. We all know that Apple's products are better than most of what's available today and that's what counts IMO.

 

[Abstract]

We're not discussing why Macs have the market share they do. The argument was that the statistics Google gathers on their visitors do not reflect the group of computer users as a whole. Unless you have evidence that there is some correlation between socioeconomic status and Google usage, this seems pretty much irrelevant.

Hey, this may be relevent: Mac users usually come from wealthier backgrounds/families than PC users, which is quite obvious considering their typical cost difference of the machines. I couldn't justify the cost until I saved up enough to buy it myself, and because Mac laptops are generally priced more competitively with Windows laptops.

I'm willing to put my testicles on the line and say that more Mac users can afford to get access to the internet when compared to PC users, and thus a larger percentage of can browse the internet because of their income. Someone who can only afford a $400 PC may not be so lucky, or may have dial-up. Of course, if they have dial-up, they may not be as keen to use the internet as often since its so darned slow. Also, someone with a fast connection can visit more sites in the same amount of time. So yes, socioeconomic status is probably relevent. Of course, I'm not a statistician or economist.

Also, gardening isn't generally a luxury that many poorer people have. Some may have access to the internet, but if you don't own space for a garden, they probably won't visit a gardening website frequently. People with a yard AND spare time can search around on Google for a gardening website. People working two jobs to support their families to make ends meet probably don't have the time to garden, if such percentages for poor people were to exist.

And the pop'n of the world is 6+ billion, not 4 billion or 7 billion. Also, I thought there were 270 million people in the US, although I'm not too sure about that one.

PS: The entire article is wrong wrong wrong. All the numbers seem so darn low. One time I was at an ATM machine when it crashed and swallowed my bank card. It then booted back up automatically.......into Winblows. Even the frickin ATM's are Winblows!! That's a lot of Windows computers, I'd say.

 

[Mav451]

I wouldn't be so hasty to generalize it that way though.

I know myself, and other college students that come from a (and i'm probably still generalizing) "wealthy or upper-middle class" county known as Montgomery County or MoCo for short.

Many kids here are spoiled with Lexuses, Mercedes, and even Acura cars without working/paying a single dime. Then they are others how bought themselves a used Tercels, Civics, or Integras b/c they were much cheaper (4-5g's and lower, the Tercels sometimes even less than 1g).

Heck, I've had cable (for the Internet) in my house for a while. But I didn't whine to my parents to buy me an expensive/overpriced Alienware, or a mediocre Dell. Nope. Instead I would use my money to BUILD an AMD system, and with this knowledge I could save money on future purchases b/c of the EXPERIENCE i gain from doing this. Oh and I save a ton of money in the process.

Who else does this? I know plenty of guys who build their own (and not just geeks either, cuz i've been approached by jocks who want to do this just for CS), and they EASILY could have asked their parents for 3,000 to get an Alienware. But they all took the other path and built it by themselves, and undoubtedly paid for most, if not ALL of the costs. $800 is manageable from even a PT job in high school, $3000 is quite a different story though.

*PS: On population? My Environmental Science textbook in 11th grade already listed the pop at 6.2 billion, and that was over 3 years ago. It is entirely plausible for it to hit 7 billion in 2-3 years, as population growth is EXPONENTIAL (and that is very bad for us )

 

[SiliconAddict]

Just throwing my .2 in for what little its worth since I've never been a numbers man.....


OK. First off this sounds like someone who is not familiar with the Windows "family" of OS's and a Mac propaganda paper. He touches on XP penetration (And oh how MS loves to penetrate people\companies\various animals OK So maybe that last one was an exaggeration.) The problem starts when dealing with Windows 2000 an OS that is essentially identical to XP. XP for all intents and purposes is nothing more then 2000 with a bunch of bells and whistles throw in. 2000 was out long enough that I can guarantee that it has a MUCH higher penetration of the Wintel platform then XP. And since the differences between the two OS's are small many users, and more importantly businesses, aren't biting on XP and sticking with Windows 2000 Professional or even Windows 98.
I think the article's author is nitpicking at this point to play the number up to look better. The 11.6% most likely includes all Mac's. Even the ones that don't have a chance in heck of running OS X. We don't have any solid figures as to what percentage of that 11.6% are old obsolete systems. (Relatively speaking of course since even a 8088 could run a word processing program if that's all you want.) Without that information we can't tell what percentage is, and could be, a viable growth from OS X. People still using OS 9 and older are there for a reason. I'm guessing either: it just works, or there is a critical app that isn't available on OS X. These users, at least for the time being, can't be depending on by apple to sustain them since they won't be buying new OS's from Apple or new apps. I suppose you could call them an island of Mac OS. Is it probable that they will upgrade in the future? Possibly, but therein lies a problem. These users have software and hardware they will have to abandon. This leaves them in a vulnerable position for jumping platforms. Unlike a user running even a current duel G4 PowerMac who is in the position of upgrade to OS X since his system already is running OS X. Again these figures aren't telling the whole story.

There is the flip side thought which also applies to PC's. Of that 240 million what can be viably upgrade to Windows 2000 or XP and what % is on a new OS such as Windows 2000 or XP? These questions are just as valid for the PC as for the Mac. Unfortunately the statistics provided above are base stats. They don't fill in the whole picture. We don't know the age of the systems that make up that 240 million or the 32 million. What if 80% of that 240 million was less then 4 years old and 80% of that 32 million was greater then 4 years old both running pre Windows 98 and pre OS X? The statistics are deceptive. When they say market share I would like the term active market share to be used. Active being the systems that can support the current or near current OS, be it Windows or OS X.


PS-Oh and if you think I'm trashing the Mac I'm not. I'm an equal opportunity OS basher. Windows, Linux, OS X, BeOS, BSD, Amiga. It's all good.

PPS- 10.4% Compaq, 9.8% Dell
I believe those stats as far as I can throw Dell's corp HQ.

 

[ebuddy889]

I pulled this from the news story at osviews, where a poster was talking about marketshare. Even more interesting to me are the claims made by a commentator, however.



I've noticed that most people about 10 years ago (around the Win 95 time) who were ready to buy a computer, usually coppied what everyone else was doing. Not having any clue what they wanted a computer for, or why to buy one. They just had to have one! because everyone else was buying them. At that time most were buying PC's off store shelves or QVC, beacuse of the prices and the amount of items you got for you're money. This still holds true.

You always see ads like complete PC system, 1.x Ghz, huge amounts of ram, tons of software and printer, cd, dvd etc for $400.00 with mail in rebate. To the uninformed economist this sounds like a steal.
So some of our Mac friends start to worry. They say "How can Apple compete with this"?

The answer is, Apple cannot. The don't even try. Apple depends on it's installed base (loyal Mac users) to buy Mac. It's that simple!

It is up to Apple's installed base (Us, we, the mac loyalists) to inform PC users that if you are finally fed up with

* All the Windows patches
* Poor quality hardware
* Confussing OS
* Viruses
* Upgrading hardware and software every 6 months to a year
* Complicated Program Installations
* Error Messages
* OS Freezes
* Incompatible Hardware
etc...
Try a Mac.

Some people firmly into PC still refuse to even take a glimps of a Mac. They believe that Macs are still in the 68k days with Localtalk networks and they claim that "Macs are way too slow for my needs on the network". They also say that "There are way more Programs for PC's than Macs". They fail to admitt that many of the PC programs they are standing firmly behind, crash at least once a day, and that crash usually takes down the entire system.

Bottom line... Apple isn't going anywhere. It's just going to take a little time.

PS Don't rely on stats and numbers.... It's like the polls in media.

 

[iMeowbot]

While the number to inherently feel to be too low on the numbers of computers, I think that we are not able to make a good estimate based on OUR surroundings.

Yes, local surroundings can be very deceiving. Just for fun, give this little toy a try.

(To save some suspense, the average worldwide annual income is about the price of an eMac.)

 

[rdowns]

Yes, local surroundings can be very deceiving. Just for fun, give this little toy a try.

(To save some suspense, the average worldwide annual income is about the price of an eMac.)

Woohoo!!!

You are in the top 0.151% richest people in the world.

There are 5,990,917,435 people poorer than you.

How do you feel about that? A bit richer we hope. Please consider donating just a small amount to help some of the poorest people in the world. Many of their lives could be improved dramatically or even saved if you donate just one hour's salary (approx $121.52)

Oh, and in case you’re interested you are the 9,082,565 richest person in the world.