Haavard

Net Applications retroactively changes stats, moves Opera to 2%

Monday, August 10, 2009 12:39:31 PM

market share, browser stats, opera, net applications, hitslink, browsers

Net Applications has retroactively changed all its statistics, and the result is that Opera's market share is listed at around 2%. This is still lower than the more accurate 3% figure reported elsewhere, but at least Opera benefits from the changes for once.

Those who have followed my blog may recall that Net Applications has previously dumped Opera's figures with no explanation what so ever.

This time, Net Applications took the time to once again admit that their statistics were skewed, and actually explained the changes:

In the past, we reported only on our raw numbers. As of August 1st, we have implemented retroactive country-level weighting in our reports. This means that we adjust our reports proportionally based on how much traffic we record from a country vs. how many internet users that country has. For example, although we have significant data from China, it is relatively small compared to the number of internet users in China. Therefore, we now weight Chinese traffic proportionally higher in our global reports. This change produces a much more accurate view of worldwide usage share statistics.

After consulting with many of the organizations we report data on, we decided to use C.I.A. data as the source of the number of internet users per country.

In addition to providing better share numbers, the reason we made this change was due to growing traffic imbalances in certain countries. Some countries were growing traffic at a much higher pace than the rest of the world and it was creating unacceptable variances in the share numbers. The reason we delayed June numbers was due to these imbalances. From now on, a single high growth country will not be able to affect the global share numbers.

The statistics from Net Applications still seem to be skewed, but neither this admission of flawed stats nor the inherent unreliability of "global" browser statistics are likely to make people think twice before referring to them as an authority in the future. That's how far "research" goes these days

Why monoculture on the Web is badMicrosoft's online Office won't support Opera or Chrome

Comments

Daniel HendrycksDanielHendrycks # Monday, August 10, 2009 2:08:42 PM

Net Applications hurt Opera a lot in the past, it's statistics said more people use Netscape then they do Opera. :S What made things worse was when Wikipedia used those states in their articles.

GroovyMicky # Monday, August 10, 2009 2:20:43 PM

More important than actual number is longterm trend. Is it falling or raising?!

FavDjiXas # Monday, August 10, 2009 2:36:23 PM

Originally posted by Micky:

More important than actual number is longterm trend. Is it falling or raising?!

Falling.

FavDjiXas # Monday, August 10, 2009 5:48:46 PM

Originally posted by WayOfTheBastard:

So why are you always spreading FUD and lies about Opera?

Hi,

September, 2008 2.01%
October, 2008 2.12%
June, 2009, 2.03%
July, 2009, 1,97%

Net App shows that it's falling

P.S. Please don't spam me via private messages next time, thanks.

FavDjiXas # Monday, August 10, 2009 8:36:49 PM

Well, I guess you can take same stats and make them look any way you want.

Unregistered user # Monday, August 10, 2009 10:54:24 PM

Anonymous writes: I am a fan of the Opera browser and have used it for years, but the NetApplications statistics clearly show that we are in a short-term decline of Opera's market share. I don't know whether that is actually true or not, but it is comments that like WayOfTheBastard's that make me be more likely to be embarrassed by my use of the Opera browser. WayOfTheBastard, I would assert that you are hurting your own cause by your tone.

Daniel HendrycksDanielHendrycks # Tuesday, August 11, 2009 1:01:23 AM

Originally posted by anonymous:

WayOfTheBastard, I would assert that you are hurting your own cause by your tone.

Truehttp://www.anchoredbygrace.com/smileys/icon_super.gif -

Mad Scientist (عادل)qlue # Tuesday, August 11, 2009 6:41:11 AM

Unregistered user # Tuesday, August 11, 2009 8:59:50 AM

Bastard writes: WayOfTheBastard: Your tone! There is no reason at all to shoot the messenger (should sound familiar). There are lies, damn lies, and statistics (should sound familiar too). Opera doesn't need your tone, please tune it down.

Nicklas Westinkilsmo # Tuesday, August 11, 2009 9:14:04 AM

This is still lower than the more accurate 3% figure reported elsewhere, but at least Opera benefits from the changes for once.

Haavard, what makes the 3% figures more accurate?

FavDjiXas # Tuesday, August 11, 2009 11:50:29 AM

Originally posted by WayOfTheBastard:

You are taking bogus stats and making bogus claims like "falling" when the fact is that the number of Opera desktop users has doubled in less than 2 years.

Why are you still lying and making claims based on the useless stats from NetApplications?

Well, I'd recommend you to read before posting.

Net Applications is NOT reporting users growth, it IS reporting market share.

Not sure why you complain even when have no idea what market share is.

Market share includes internet users growth. If there are 100 users, 10 are using opera, it's 10% market share for opera. Lets say 100 more people get internet = 200 users total and 15 are now using opera. That's 7.5% market share for Opera = decreased, even if it has more users.

Charles SchlossChas4 # Tuesday, August 11, 2009 9:33:57 PM

Net Applications changed there ways

I wonder where Asa is? I know he wants to comment

53north # Wednesday, August 12, 2009 4:06:01 PM

I'd have thought opera will rise exponentially as Wap sites like peperonity become opera au fait.
Sniffing Wap and giving them links to opera mobile and mini could bring in many eastern euro and central asian users.
It wouldn't be impossible to charge for an enhanced opera mini 5 - leave a hybrid 3 or 4 version free. Asking for €3 of phone credit for something that's infinitely more useful than a game or ring tone seems fair dinkum?

Haavardhaavard # Wednesday, August 12, 2009 4:42:24 PM

Originally posted by kilsmo:

Haavard, what makes the 3% figures more accurate?

Internet World Stats reports on the total number of people online, and we report to the public how many users Opera has every now and then. Opera's 40 million users at the time was about 3% of the total number of desktop users reported by IWS at the same time, if I recall correctly.

Tolque53Tolque53 # Friday, August 14, 2009 11:54:15 AM

Well, I also think - looking at my friends - that Operas market share shrinks. Just because many have experienced better surfing (pages work) using other browsers and do not use the sophisticated functionality opera offers.
I personally still use it, but I seldom use it to use Facebook, for example (even in 10 beta 3 now), because its way too slow, especially when using apps.
Still, Opera is my standard browser, and Opera Mail is my standard email software - that has now been drastically improved.

Khaled KhalilKhaled-Khalil # Sunday, August 16, 2009 5:24:24 PM

Originally posted by haavard:

and we report to the public how many users Opera has every now and then. Opera's 40 million users at the time[...]

and how does opera get the real number of its users ?
number of downloads for example don't tell anything, many people download softwares just to try it, many download alike shareware from third party web sites, or even copy it from friends offline by storage media.

Nicklas Westinkilsmo # Sunday, August 16, 2009 6:03:19 PM

I would argue that even Opera can't get the real number of its users. (for the same reason that Mozilla can't either)
You can use the update mechanism to track the number of used installations, but:

* A user can do several installations on different machines. (even several installations on the same machine)
* The browser can be used by several people in a family.
* The user will maybe use a number of other browsers too.
* To get to 100% in total, you can't really count all the users of a browser, but will have to count the amount of data from the different browsers to decide which browser the user is using the most. A browser maker does not have this information.

Haavardhaavard # Monday, August 17, 2009 12:48:14 PM

Yes, the update mechanism can be used, and is being used, to calculate the number of users.

Opera's user numbers are the number of active users, not data or anything like that. Actual data usage is only reported for Opera Mini.

Khaled KhalilKhaled-Khalil # Monday, August 17, 2009 10:41:37 PM

well, opera mini usage (and terbo's) can easily be tracked, but i am sorry i can't accept any estimation through updates, i even find net application or any other flawed stats much more reliable than updates results.

Haavardhaavard # Tuesday, August 18, 2009 8:08:22 AM

Then I'm afraid you have misunderstood. Actual user numbers based on automatic updates are much more reliable than browser statistics. There's no question what so ever about that.

While browser statistics rely on questionable samples and sample sizes, counting the number of users based on update checks gives you a fairly straightforward data set to work with. Potential errors sources are either easily taken into the calculation, or don't really make a difference, depending how the calculations are made.

53north # Tuesday, August 18, 2009 9:11:01 AM

I noticed on Alexa's global standings, my.opera has gone up from 723 a month ago to 702 last week. (Google is no.1, Yahoo 2 etc..). Anything lower than a 1000 is v. good and an important link.
I believe the standings show importance to the net, through weight of traffic and quality of content.

Cutting Spoonhellspork # Monday, August 24, 2009 3:46:06 AM

Browser.js and several other routinely fetched files create a fascinating measure of the usage pattern. If Opera ASA can look at a daily report of how many "browser start" events were issued, this provides a crude - though compelling - evidence as to how often people use Opera. Crossing this against the known install base, they might further estimate how often the average user reaches for Opera to do work.

If My.Opera is growing, good. Link and the phone versions have likely done wonders to increase traffic. Do synch events register as hits to My.Opera? Even if not, Link has been leading users to My.Opera, and lack of invasive ads has led some users to recommend the service to others.

So Opera can draw on update checks and startup events, total hits and number of users on My.Opera, very accurate numbers for Mini and Turbo (both users and traffic). And that would still only be an internal statistic. I couldn't care less about the rest of the internet, so long as Opera is profitable and still maintains a desktop version.

Haavardhaavard # Monday, August 24, 2009 6:49:44 AM

Link, Unite, etc. users are not counted as My Opera members until they actually log in at my.opera.com.

Cutting Spoonhellspork # Thursday, August 27, 2009 3:59:08 AM

Really? I like that. Thank you for clarifying.

illiad # Thursday, November 5, 2009 2:04:07 PM

good topic

how about releasing these numbers, in a 'trend' graph, to 'show up' the competition??

Correct me if wrong, but Alexa only track what sites are accessed, not what browser is used to access them?? If you check for 'microsoft' it shows that most are not interested in going there, I guess due to the *total* lack of a support forum! The big ones, of course, are email sites, not even google or yahoo..