Compete Toolbar Bug With Google Reader – Compete vs Alexa vs Quantcast Stats

I have had long term problems with the Compete Toolbar as I pointed out some time ago on my friend Paul’s blog, I think I actually made that comment on my first visit to his blog.

The only time I actually test the Compete Toolbar is when I get an upgrade for it. Firefox detects a new upgrade, downloads it, and switches the toolbar on automatically from its default state on my computer – off.

The reason I keep it switched off is because with the Compete Toolbar switched on, Google Reader doesn’t work. It has been a good 3 months since Paul noted a problem. I just downloaded a new version, and again it doesn’t work with Google Reader on my system.

Compete Statistics

I have been disappointed with Compete Statistics for some time, especially within various niches that are technology related.

I decided to do a little roundup of blogs in the SEO SEM & Make Money Niche – no disrespect to anyone I missed out, page load time is going to suck anyway with all these images. Some sites I checked out and they fell into line with others which have similar Alexa traffic, so I just kept one example.

Problogger.net

Current Alexa Rating:- 2309

Quantcast Problogger

problogger

Shoemoney.com

Current Alexa Rating:- 1281

Quantcast Shoemoney

Shoemoney

Graywolf

Current Alexa Rating:- 4993

Quantcast Graywolf

Wolf-howl

Pronet Advertising

Current Alexa Rating:- 10015

Quantcast Pronet Advertising

Pronet Advertising

Search Engine Journal

Current Alexa Rating:- 4264

Quantcast SearchEngineJournal

Search Engine Journal

John Battelle’s Search Blog

Current Alexa Rating:- 7684

Quantcast Batelle

Battelle

Search Engine Watch

Current Alexa Rating:- 957

Quantcast SearchEngineWatch

Search Engine Watch

TopRank Blog

Current Alexa Rating:- 7610

Quantcast TopRankBlog

TopRankBlog

Andy Beard

Current Alexa Rating:- 32K going on 20k

Quantcast AndyBeard

AndyBeard

Statistics Analysis

The Alexa stats just don’t tie up with the Compete stats, even in a similar niche. Why do SEJ and SEW do so much better than Problogger and Shoemoney on Compete scores when both have very high (and legitimate) Alexa scores?
John Battelle stats are fascinating – he has a huge number of subscribers based on his recent stats from a few days ago after Google Reader was included but his Alexa score and number of comments doesn’t match. John seems to have a different kind of reader.

Do corporate intranets and default feeds on start pages (before you sign up) get somehow added to subscriber data?

I have always looked on the number of comments as being a good indication for blogs as to whether Alexa data is consistant, and generally within this niche it is, with the exception of John Battelles stats. I can think of reasons for that, as he is a “trusted” news source.

It should also be noted that Compete is a relatively new service, younger than all of the sites listed. I didn’t include SearchEngineLand specifically because it is a new site and thus might not have been a fair comparison.

As for my own data… wow, I am in the top 1 million now for Compete.

I feel much more inclined to use Alexa data as an indicator over Compete, not only because I can actually use it without breaking Google Reader, but also because the statistics for Alexa at least seem to have some level of consistency.

Time to close Firefox to switch Compete off again until the next update.

Update: I have now added Quantcast figures

Quantcast

Quantcast is a strange entity, in that the figures for traffic can be highly accurate. I know the figures they have for my site are 100% accurate. The reason they have exact data for my site is because I have been running their tracking code on my site for the last month and have my site registered with them.

For sites that aren’t registered and running the tracking code, the data is fairly weak – I am not sure where they get the data from, but it seems to be fairly dubious.

Its another piece of code you end up placing in the footer of your site, and I am sure the tin-foil hat brigade are not too happy about sharing everything about their sites, but it seems Quantcast get the data even if you don’t use their tracking code. It is probably best to register and then decide which data to display.
Lots of data isn’t always displayed for lower ranked sites – that certainly seems to be the case with this site.

From the Quantcast FAQ

How can I fix my site profile?

There is a sure-fire way to improve the accuracy of these numbers: by placing a measurement pixel on your site. Add a pixel and let our technology do the rest. Within just a few days you’ll see more refined, accurate and representative numbers for your site’s audience.

How accurate their ratings are for sites that experience a growth in traffic are unknown, and the drops in reported traffic on some sites are strange. Where you host seems to make a difference with how much data they have if you don’t have the code installed.

Update2: Conversion Rater has been doing a similar analysis trying to create a rankings table based on this data. Unfortunately it isn’t taking into account why some of the metrics data might be totally inaccurate, such as sites not registered with Quantcast.
Clarification: Pat does mention screwy data for sites not including Quantcast code, but the way he says it suggests that this is affecting those on his list that have a 15,000,000 rating. Of the 16 sites on the list, only 4 might have registered with Quantcast. It is just not worth using the figure. Comepete seems to be all over the place as well, with 6 sites not having any data at all.

Liked this post? Follow this blog to get more. Follow

Comments

  1. says

    I have looked at it, but they might be using some of the same data sources as Compete.

    I am going to add Quantcast figures to the list for all the same sites.

  2. says

    Hi Andy,

    Cool analysis of your own. I agree that some of the Compete and Quantcast data is screwy, especially if you aren’t “quantified” by Quantcast, and although in your blog post you say I don’t account for that, I actually do mention in the article that there are problems with that data.

    Also, because I’m using that data combined with Alexa, Technorati, Pagerank, and Yahoo links I think the overall picture becomes pretty accurate to the popularity and influence of these blogs

  3. says

    Hi Pat, you mention it as a factor for people with no data, but not for the people with what I would class as “screwed” data. Quantcast and Compete have data for Shoemoney, but it is seriously messed up.
    Some of the sites I can even grab their keyword data (I am sure Graywolf will hate that) because they aren’t registered.

    These sites don’t use the Quantcast code, but they do have figures for them… that are totally wrong – it is not the same as the ones in your raw data with 15,000,000

    Technorati can be gamed just by being in a Blog network and lots of other ways.

    Yahoo links include nofollow links, forum, blog comments even if nofollowed, sigs etc etc

Trackbacks