In the social media sphere, most people use Compete.com as a nice standard to evaluate the popularity of a site. Sure, there’s Alexa and a few other sites out there that measure traffic too, but lately I’ve mostly been hearing people talk about Compete.com traffic instead of the other ones.
Quantcast offers more accurate data because it is actually based on scripts that is embedded in these sites, but because of that, the sites it covers are also limited.
How Compete.com evaluates traffic
Compete.com evaluates each site’s traffic by a sample group of roughly 2 million internet users (what they say is 2% of all internet users) based on ISPs, ASPs, Opt-In Panels and the Compete.com Toolbar.
Through some complex statistical calculations, they roughly estimate how many people really go on your site from this sample group.
I would say Compete.com gives a fairly good feel of if a website is popular or not, especially if you use it to compare to other sites in the same field (hence the name “Compete.com”).
Since it uses ISPs (Internet Service Providers) as one of their metrics, it is stronger than sites like Alexa that mostly base traffic on those who use their toolbars.
However, there’s still a lot of inaccuracy in Compete.com. After years of usage and comparing it to Google Analytics data from many sites that I managed, I have some interesting observations of how Compete.com’s behavior differ from actual traffic. (Note: Google Analytics does not reflect 100% actual traffic, but it is definitely a whole lot closer than Compete.com, so I will take that number as a standard).
Compete.com does not measure traffic outside the US
One thing to quickly establish about Compete.com is that it only measures traffic from the United States. That’s because its ISP partners are in the States, and it has a much bigger penetration in terms of US ASP (Application Service Providers) from other websites.
If you look at a very popular site in Taiwan, Wretch.cc, you will see that Compete does not report any traffic from them.
However, almost all my friends from Taiwan has a Wretch account, and it’s pretty much as popular as Facebook there. I am 99.99% certain that it at least has hundreds of thousands of monthly users.
Google recognizes the popularity of Wretch and gives it a Page Rank of 7/10, an astonishing high number equal to the blogs of Guy Kawasaki and Mashable.
From this you can see that it is NEVER a good idea to take Compete data and tell someone that an overseas service is not very popular.
Compete.com inflates traffic for sites in the Social Media World
If you are looking at a website that is popular in the social media industry, then Compete.com traffic will likely look better than it actually is.
This is because there are way more people in Social Media that uses the Compete.com Toolbar. If 10% of the people with Compete.com Toolbars visit a certain site, Compete.com will come to think that 10% of internet users go onto this site (before factoring in all the other metrics), and they will help boost your overall score on compete.com.
Also, I am not sure who are the Application Service Providers of Compete.com, but I’m guessing that it has something to do with various popular websites, and visitors of those websites would also get extra “Compete.com Points” for the sites they visit.
Take my blog for example. Compete.com rated my traffic last month as roughly 21,000 unique visitors. However, that is much higher than my actual traffic.
Compete.com mostly thinks that my site traffic is that high because many of my readers have the Compete.com Toolbar, and that a lot of people find me through Social Media sites like Twitter or other blogs. Again, I’m not 100% sure how these sites relate to Compete.com, but it seems to me that they are weighted more on Compete.com than the actual traffic.
Compete.com under-ranks Google Search traffic
This one is pretty interesting, and perhaps the biggest insight this post has to offer. I have noticed time and time again that sites that have huge Google search traffic is underrated…by a lot, on Compete.com.
Lets take Sean Percival, a respected figure in the online content and SEO world. In one of his blogposts, Google Analytics showed that he had close to 140,000 unique visits on his site that month, of which 105,000 of them are from Google.
However, Compete.com only shows that his traffic is lower than 30,000 during the same month. You can see that Sean’s actual traffic is WAY over what Compete.com thinks he has. Interestingly enough, if you take out all his Google traffic, compete.com would be quite close to the actual number.
Another example would be Steve Wiideman, a credible SEO Expert who is also on FD’s advisory board. Yes, many people call themselves SEO Experts, but he is truly one. You can tell by Googling “SEO Expert” and “SEM Expert”.
Even though his site consistently ranks as the first or second results for SEO Expert and SEM Expert, terms that are searched quite a bit, his compete.com traffic is lower than 3,000. It is pretty convincing that his real traffic is way above 3,000 a month, and we can conclude that Compete.com underestimates his traffic too.
How to tell if compete.com is underestimating a site’s traffic
I’ve made the mistake of prejudging a website’s popular by looking at the compete.com traffic. I have even made a fool of myself before because of that. To avoid these mistakes here are a few tips to go by:
- If a website looks like it could mostly be used by foreign users outside the US, assume that the traffic is way too low.
- If a website has high Page Rank (you can check PR easily with various Firefox Plugins) but low traffic on Compete.com, assume that the traffic is underestimated.
- If you are on a blog or social media site and it has SOME compete.com traffic, assume that it is ovestestimated.
- No matter what you do, don’t take compete.com data and report it to others as a fact. Always take Quantcast.com traffic over compete.com if they happen to have data on the site.
Ways to boost your compete.com traffic
Here are some pretty straight forward tips to help your compete.com Traffic look better. Most of these techniques are standard ways to help you get real traffic too, but having compete.com in mind. Yes, having high traffic on compete.com is not nearly as good as actual traffic, but sometimes you will get more opportunities when people believe that a lot of others like your site.
- Download the compete.com Toolbar, and tell all your friends/users/readers to download the compete.com Toolbar.
- Learn how to generate traffic from various social media sites. This is a professional art that companies are paying to do, but once you know how to establish a brand name in these sites and drive traffic to your own site through it, compete.com seems to value that more.
- Get publicity through blogging and get publicized by other blogs. Again, a good amount of compete.com Toolbar users read blogs, so try to capture a larger portion from that demographics.
- Even though Search Engine traffic does not rank well in Compete, when you are optimizing your search engine strategies, focus on search traffic that will come from the States.
Good luck on getting a lot more actual traffic on your sites! If you are a blogger, the Viralogy Blog Rank might help!
Would you like Yu-kai Chou to work with your Organization?
If you are interested in working with Yu-kai Chou for a business project, workshop, speech or presentation, or licensing deal, please fill out the form below.
10 thoughts on “Compete.com Traffic vs Actual Traffic”
Hey! Thanks for the encouragement! Data is such an important part in decision making, but people need to know where did the data come from and how accurate it is. Same with surveys and such.
Btw, I checked out your site. The concept sounds awesome and has some relationships with what we want to do with Viralogy, but I didn’t get how to use it at this point. Is it still in private beta mode? There might be opportunities to work together one day:)
Found this article because you followed me on Twitter. Very glad I drilled in to read it. Great analysis and your findings match what I’m seeing as well. We’re working on a new social media monitoring tool and have basically abandoned compete.com data because of the limitations you describe above.
We also looked at Alexa and its results are laughably poor for blogs and such where a blog may have quite a loyal and active following of people who are smart enough not to have the Alexa toolbar installed.
This is a very tough problem to solve and can probably only be addressed adequately by a major search provider. Hopefully Pagerank will evolve into a more open standard at some point.
Thanks! You are awesome! 😀
Loved your article! very informative thanks! i will definitely retweet this 🙂
Yea! That’s why one should not use Compete.com for any sort of decision-making judgements, but use it as an evaluation tool.
This is very interesting. I can see where it can become frustrating for a marketer if clients use Compete as their traffic barometer…I guess the key is education. Thanks for the insight!
Haha, inaccurate is usually more preferable than being bankrupted 🙂
Great post! Thanks for sharing such a good information! Maybe it’s time to switch to Nielson Online! 🙂
Thanks for providing the very nice information.
Hmmm, to spam or not to spam. That is the question.