90:9:1 – the odd ratio that technology keeps creating

http://www.theguardian.com/technology/2015/dec/12/ratio-technology-mozilla-firefox-os-90-9-1

by andyjohnson0 • 11 years ago

137 51 11 years ago

I'm not familiar with the HN API, but is it possible/can someone come up with submitters:commenters:inactive accounts for say the past year to see if it comes out close to 1:9:90? Does the "1% rule" seem to apply to HN?

tacos2• 11 years ago

Any site that you find interesting likely has an element or structure that causes it to skew that pattern. Wikipedia doesn't follow that ratio, for example. Also the metrics around social media stuff are so bizarre (and so easily manipulated) that it's hard to get decent data. Trump's engagement metrics on Twitter are in the fractions of a percent. CEO posts on LinkedIn get a few hundred reads and four comments. So is that 100:1? Or is 400 reads versus a million followers the important ratio? It's all over the place and you can pluck whatever curve you want by selecting datasets and ignoring important outside influences such as LinkedIn's idiotic design or Trump's pacing of Tweets.

GavinMcG• 11 years ago

I think the article describes exactly the two factors in play:

The more you look for it in networked environments, the more frequently some sort of 90:9:1 ratio seems to emerge once the market matures.

First, of course something will show up the more you look for it.

Second, "once the market matures" is completely up in the air. Was the market immature in November 2009, when IE had 65%, Firefox 24%, and Safari, Chrome, and Opera each 2-4% of the browser market? Is it mature now? I don't think "immature" is the word that comes to mind to describe the browser market, for many. Yet there's certainly no 90:9:1 ratio in browser share currently.

notahacker• 11 years ago

In fact, with browsers the 90:9:1 ratio was present around the turn of the century when the internet was much more immature and the trend has been away from that.

dogma1138• 11 years ago

I think mature isn't the right word I would use stable instead.

10 to 1 ratio also pops up in business quite a bit between leaders and competitors in high entry barrier fields. If your competitor is 10 times bigger than you then 90:9:1 ratio pops out quite naturally in many places espcially with even the slightest selection bias.

Pretty much most Leader:challenger:niche selections will maintain a similar ratio, just your run of the mill power distribution.

mikeash• 11 years ago

"Once the market matures" seems to make this a tautology. Oh, the ratio isn't 90:9:1? Must not have matured yet.

fragsworth• 11 years ago

There are plenty of technologies/companies that have no clear 90:9:1 ratio. Databases, programming languages, consoles, portable consoles, laptops, games, game developers.

The "90:9:1 ratio" is a specific case of a zipf curve. Popularity is always fluctuating, but generally follows a zipf curve for distribution.

coryl• 11 years ago

Isn't it better/easier to measure maturity by looking at a product's major generational milestones or releases? ie. the 4th iPhone, or the 4th Tesla. Does that help? Most new things seem to follow that growth curve of early adopters, mass market, etc.

sparkie• 11 years ago

"The market matures" in my view means "innovation has stagnated". The only apparent differences between the competing products are cost, convenience and the network effect. The technological differences are uninteresting. It seems in this environment, the smaller players attempt to stay alive by becoming replicas of the market leader.

cmarschner2• 11 years ago

Technically the browser "market" is not a market, because browsers are for free. That's very unusual.

Hence whoever invests most in its development + distribution can gain market share. Similar to Android, Chrome is strategically important for Google to keep users in their ecosystem, so it makes sense for them to overinvest. It's just a loss leader paid for by advertising dollars. Also distribution is cheap for them through google.com.

Microsoft has kept IE in the Windows org, and for the longest time did not invest enough to pay down technical debt. For the longest time investments couldn't be justified with advertisement dollars, so after they "won" the browser war with Netscape's demise they cut down much of the IE team.

Firefox devs paid down Netscape's technical debt for free to escape the IE monopoly. But distribution is a problem for them since they neither own the desktop OS nor the web's primary entry point.

iopq• 11 years ago

It's just a loss leader

I think having Google bundled makes it a clear winner. If Firefox can pay its developers with just the search engine bundle, and it's less popular, it would probably stand to reason that Chrome has paid off much more.

In a world where Firefox is the #1 alternative to IE and Chrome has very weak showing Firefox either asks Google for a ton more money or switches to Yahoo and Google loses a lot more market share.

kami8845• 11 years ago

It's still a market if every product is free. And they're not even really free, you pay with potential advertising revenue, vendor lock-in and with data that you generate.

Gladdyu• 11 years ago

Welcome to the power law (https://en.wikipedia.org/wiki/Power_law)

uxhacker• 11 years ago

Also Zipf law. http://www.hpl.hp.com/research/idl/papers/ranking/adamicglot...

tacos2• 11 years ago

The article fails to provide a single example where those are the specific ratios (versus arbitrary curves), provides a number of counterexamples, then concludes "proven!" To wash off the stupid, I found myself re-reading this excellent primer by Ben Golub (Harvard Dept. of Economics):

https://www.quora.com/In-what-conditions-would-you-expect-a-...

tylermauthe• 11 years ago

Now look at desktop OS sales: the ratio stands in the most recent quarter at about 91:8:1 between Microsoft’s Windows, Apple’s Mac OSX, and “self-build” machines which probably get Linux.

This is a drastic oversimplification... Some of those Windows machines will have Linux put on and many "self-build" machines will have Windows (or Hackintosh OSX). The vast majority of the Windows and OSX sales will likely stay as Windows or OSX boxes, so the data is still largely valid - but I feel this particular measure is a bit of a reach.

Furthermore, we can say this is isolated to technology, but in my opinion this happens in Facial Tissue sales just as often as tech. This is a law of human preference: on average 90% of people follow the herd, 9% go for a close alternative, 1% go hardcore.

This is even true of characters in TV shows! 90% of people favour the hero, 9% are into the side-kick, 1% dig deep into some random character, possibly even creating new backstories.

[edit: clarity]

xyzzy4• 11 years ago

Looks like somebody rediscovered how power law distributions are prevalent in so many different statistics.

SixSigma• 11 years ago

Pareto ?

Robert Kiyosaki suggests 80/20 is an average but in the world of money, the most suitable rule is 90/10.

http://1-million-dollar-blog.com/what-is-9010-rule-of-money/

astazangasta• 11 years ago

While this is insipid, the tendency of the Internet to a winner-take-all system is interesting and needs more commentary. Is it the product of the massive consolidation of wealth and corporate power? Or is it a feature of the Internet? Should we avoid it, or mitigate its effects (monopoly power)?

networkeffect• 11 years ago

Network effects are called network effects because they predate the internet. They are well-studied, and those who from them are adept at using the political system to erect and maintain barriers to entry for challengers.

nabla9• 11 years ago

This is caused by network externalities. Both direct and indirect network externalities. When a product’s or platforms value to the user, third party developer, or hardware manufacturer increases as the number of users of the product grows, this creates strong positive feedback.

Open standards and monopoly laws can reduce this externality and keep them from becoming monopoly. Android can call iPhone, and software can be ported from one platform to another.

tedmiston2• 11 years ago

85:14:1 for Android:iOS:Windows Phone might be right globally, but... that's surely skewed by developing countries.

My gut feelings say iOS would be very dominant, but it'd be interesting to see how the ratios compare in startup hubs to the rest of the U.S. Or, even better, broken down by income level, state, or some other useful demographic info.

Edit: Sounds like I should do some more research on Germany :).

TillE• 11 years ago

Android is hugely dominant in Germany and most of the rest of Europe. iOS is somewhat higher than 14%, but still, the US is a clear outlier in its relatively even mobile OS numbers.

digi_owl2• 11 years ago

I dunno. Fire up a map of Europe on Statcounter and you see Android everywhere but UK, Scandinavia, and maybe Netherlands. The rest has Android sitting on top (and i would hardly consider Germany a "developing nation").

aembleton• 11 years ago

The UK is pretty much neck and neck when it comes to Android vs ios. If you change the map to just this month then Android comes out ahead: http://gs.statcounter.com/#mobile_os-eu-monthly-201512-20151...

tedmiston2• 11 years ago

Wow, their data is enlightening.

Sept to Nov 2015* by Mobile OS Worldwide http://gs.statcounter.com/#mobile_os-ww-monthly-201509-20151...

These percentages are from the CSV export:

  Country  Android  iOS    WP
  -------  -------  -----  ----
  U.S.     46.81    50.7   1.63
  Germany  67.72    28.34  2.67
  UK       43.14    44.2   3.62

  European countries dominated by iOS

  - Denmark
  - Faroe Islands
  - Gibraltar
  - Guernsey
  - Isle of Man
  - Jersey
  - Liechtenstein
  - Luxembourg
  - Monaco
  - Norway
  - Sweden
  - Switzerland
  - United Kingdom (though 44.2% iOS vs. 43.14% Android)

  Asian countries dominated by iOS

  - Christmas Island
  - Japan
  - Macao

If you filter out the rich/small/weirder island nations, both lists are... quite short.

A few more for the curious:

  Top 5 Android Countries by %:

  Country      %
  -----------  -----
  Myanmar      92.32
  Niue 	       87.05
  Iran 	       86.86
  North Korea  86.53
  Poland       86.31

  Top 5 iOS Countries by %:

  Country           %
  ----------------  -----
  Norfolk Island    77.27
  Christmas Island  71.74
  Monaco            70.07
  Japan             65.4
  Guernsey          62.04

*The free version limits to 3 months for a map.

tallanvor• 11 years ago

Poland is very interesting... Android is the clear winner with Windows coming in second with over 7%. iOS is under 2%.

curun1r• 11 years ago

The other thing to consider is the normal consequence of the 90:9:1 rule, which I'm not sure applies to the current mobile market. One of the reasons why that ratio is problematic is that one company is reaping the vast majority of the profits and can invest more heavily than competitors can. This leads to stagnation as the entrenched company invests just enough to keep their product superior to competitors and keeps the rest as profits.

But the mobile market is unique in that the revenue share is much more even. Apple's share of mobile profits is far more than the 14% of mobile users mentioned. This means that despite Google having a large lead in the overall number of users, both companies are able to invest similarly in their platform and the competition between the two is far more intense than almost any other 90:9:1 situation. The same holds true with the ecosystems around the mobile platforms...iOS users are much more valuable to app developers and the Android market is much more fragmented which means that both ecosystems have a thriving developer community.

hodgesrm• 11 years ago

While this article cites browsers as an example of 90:9:1 curiously I use 3 of them, generally simultaneously:

Chrome - For work

Safari - For personal stuff

Firefox - For occasional stuff that falls on the other two browsers (sadly more common than one would like)

In the case of browsers there is a huge amount of context related to things like work and private online accounts, e.g., at Google. Using different browsers is a simple way to keep your personalities separate.

ansgri2• 11 years ago

I agree. Browser vendors should recognize this to eat up the remaining percents. It should be easy for me to have several fully isolated Chrome instances with different desktop icons and clearly distinct windows (this is where UI color-theming shines).

phamilton• 11 years ago

Chrome user profiles does the trick for me.

leppr• 11 years ago

I don't know about Chrome but it's very easy to do with Firefox, just start it with "-P", optionally followed by a profile's name to bypass the profile manager window.

ebbv• 11 years ago

This is just one of those horse shit "rules" that you can find a lot of places if you're willing to look for it and do some really liberal rounding.

cbr• 11 years ago

"the only product that has ever been advertised on Google’s front page"

The Nexus One and 7 were too:

* http://searchengineland.com/surprise-googles-home-page-promo...

* http://www.cnet.com/news/googles-nexus-7-tablet-pops-up-in-r...

plank• 11 years ago

The examples in the article could just as well be examples of the pareto principle.... ( https://en.wikipedia.org/wiki/Pareto_principle )

Kiro• 11 years ago

OT but is anyone else seeing a Zalgo font? Looks really broken.

hackuser• 11 years ago

I wonder how well it describes participation in politics.: It rings true: Only 1% really thinking about the world and creating new ideas, 10% thinking critically about what the 1% say (but letting the 1% define their options), and 90% following the herd.

mrcactu5• 11 years ago

Is this just the geometric series? 1, 9, 9×9 = 81

The other ratios in the article are similar: 85:14:1 or 91:8:1

This is just basic Social Network Analysis (https://www.coursera.org/course/sna)

This rule has two parts, given a choice of selections

    A: there are only a few "key players" 
    even if the original selection was very large

    B: the ratio of #1 to #2 is about the same as for #2 to #3

This is why multiple choice always have 3 choices typically - we do not even notice #4...

- - - - - - - - - - - - - - -

loosely related:

Viggo Brun wrote about algorithms to find to simple ratios... he was trying to tune a piano!

He calculated the ideal tuning ratios are close to some simple fractions

    log(2/1):log(3/2):log(4/3) ≈ 12:7:5 ≈ 53:31:22

http://retro.seals.ch/digbib/view2?pid=ens-001:1964:10::21

rd108• 11 years ago

Power laws show up in many systems, as the 80-20 Pareto rule, graph theory's "scale-free network", even electromagnetism and gravity.

Double_Cast• 11 years ago

Sometimes, things differ by orders of magnitude. News at 11.

http://www.smbc-comics.com/index.php?id=3777

90:9:1 – the odd ratio that technology keeps creating

Related Stories