We get this question all the time, so I’m going to answer it here and then point people back to this post.
Alexa is the most well-known of the traffic measurement services, but also the least accurate. Let me explain why. Alexa rankings are based on the behavior of the folks who have downloaded the toolbar. No wonder that tech-heavy and venture capitalist sites have fantastic Alexa ratings– the tech people are disproportionately represented as Alexa users. The Alexa on this blog fluctuates between 50,000 and 250,000, which is better than the 30 million plus sites being tracked. Though my blog gets only a few hundred folks a day, it has an Alexa rank better than many sites that get 10,000 unique a day, such as some of our clients.
What you need to know about Alexa is this:
- If the Alexa rank of a site you’re looking at is not under 100,000, then the odds are the traffic is low. And if your traffic is low, then the activity of just a few users can make your number bounce from 2 million to 20 million– it’s just noise. Remember, Alexa estimates your popularity by extrapolating the behavior of a few users up to the entire web. But the Alexa userbase (those who use the toolbar) is highly skewed.
- You can game the Alexa toolbar by having your friends visit your site or employing a bot service– seems silly to go to that effort just to better your rank. I suppose geeks can have their boob jobs, too.
- You CAN compare Alexa figures between sites in the same category– so one daddy blogger versus another daddy blogger is fine since they’re likely to have the same mix of Alexa users. But a tech blogger versus a kids’ site? Nope.
Compete and Quantcast are usually far more accurate. Quantcast allows you to be “quantified” by placing a tracking pixel, so it’s the most accurate. Yet I don’t see the benefit in sharing your stats with the world– but then again, I don’t understand why some people share some of their most personal details in public forums. When we do analysis, we use multiple third-party traffic sources– and typically, Compete and Quantcast go together, while Alexa deviates. Some thoughts about these two services:
- Compete is great if you want to compare multiple sites at the same time– hence, the name “compete”. I haven’t personally paid the cash to use the Pro service, but can’t imagine that any paid service is really worthwhile, given how many free tools are available.
- Quantcast is great for individual site demographics and tells you what other sites are in the same category. I’d ignore the education and income measures– no way they can accurately determine that. A neat trick is to take the related sites to spit out by Quantcast (your competitors, I’m assuming) and then paste those domains into a spyfu.com or other keyword research tool to get their keywords.
ComScore is for big companies that have a lot of money to spend on reporting, although the unwritten rule is that if you pay ComScore for their reporting, you might somehow rank better in their various reports. ComScore doesn’t matter to you unless you’re VC funded or a big company that is depending upon who ComScore says is top in a particular category for your valuation or ad rates. There is a minor conflict of interest here, but not much more so than the fact that automobile review magazines accept ads from car dealerships, or that Yelp takes ads from restaurants that have reviews in their directory.
But the most important tool to use in grading your site is your own analytics– and you should be using Google Analytics here unless you are an enterprise client that can afford a custom clickstream tool to do click-level attribution, or are perhaps doing some crazy form of lead generation.
The question is what your goal is with these 3rd party traffic tools. And there are different tools for different jobs depending on whether your aim is keyword research on competitors, exploring new demographics for a product you want to launch, researching the other players to demonstrate where you’ve kicked ass over the last year (to get a raise), or perhaps it’s just a nice stat to consistently track each month over time.
So what tools do you prefer to use for assessing popularity on the web? My favorite is still Alexa because it’s easy to use and commonly understood.