Ever wonder how much traffic a Web site is getting?
We’ve already talked about how tough it is to get accurate stats. Alexa, the favorite of many people because it offers lots of easy-to-get free data, is flawed. And new competitors, such as Compete and Quantcast, which claim to offer something better, also have drawbacks.
Traffick has a good summary of the providers, but still leaves you banging your head against the wall in frustration that there’s no easy, reliable way to do this:
Tags: co:Alexa, co:Compete, co:QuantcastBased on the evidence I’ve sifted through, there’s not a shred to suggest that Compete.com is better [than Alexa.com] at this stage, and some to suggest it’s actually worse. That shouldn’t be surprising. Going strictly on toolbar installs, the upstart service is bound to have less data, and less representative data. They try to augment that with “ISP relationships,” I gather, a practice which is quite old and generally not well explained. (Are Hitwise’s ISP relationships better than Compete’s? Are Wordtracker’s better than KeyWord Discovery’s? Does anyone do anything transparently in this industry?)
The CEO of Quantcast, Konrad Feldman, was good enough to talk to me a couple of months ago. I’m still uncertain about where the service is headed, given its newness. Essentially they attempt to combine various data sources to arrive at more accurate rankings, including demographic information for a site. To aid this, sites are encouraged to download some javascript code. Interesting, but no different from the scores of competing analytics services hoping to do the same. Who can get every site or even a lot of sites to install that code?
28 Comments
-
Lauren Moores said:
Hi Matt- It is good to see the discussion about who has the best metric out there, especially when Compete is included. Our ‘fledgling’ company is 6 years old with a methodology that has been proven with our B2B customers. We do not rely only on our toolbar installs; the toolbar is a recent addition to allow for greater sample. We collect and license our data to allow for better distribution of the actual population. We believe our sample of 2 million+ consumers is the biggest US panel available. This panel size and our proprietary normalization process allows us to more acccurately reflect consumer behavior than our competitors. But …we love the feedback because we are always looking to make the data better. Thanks! Lauren
-
Webanalyticsbook said:
How about giving Compete and others more time? Manual trackback: http://www.webanalyticsbook.com/archives/520
-
Konrad said:
Matt, We plan to alleviate the frustration with a participatory open model. You’re right, any estimation method will have its drawbacks and panels are no match for the diversity and fragmentation of the Internet. We launched the Quantified Publisher program to incorporate measurement pixels in our models. This program now serves 1 billion pixels for as many as 20 million visitors each and every day. By combining pixel data with a traditional panel approach we believe we have created a convergent model that allows direct publisher participation and will continue to improve as the program expands.
-
bobby said:
look at http://www.enquisite.com
-
David said:
Are we discussing “free” traffic metric sources? Larger websites that need third party validation of their numbers (or competitors) do not even look at the Alexa - Comscore, Nielson, Hitwise and the like are the sources that are used - although the first 2 can have trouble with newer or lower traffic sites.
It’s crazy how hooked on Alexa some people can be…even after they see all of the articles disputing the numbers how easy it is to game the system.
-
Andrew Holt said:
Interesting post Matt - as free sources of data, I’m very interested in seeing how Compete and Quantcast improve. I believe all of the analytics companies (free or paid) will have some drawbacks, but all also offer their own advantages. We may start offering all three of the major free services as data points on our service…
-
john hartley said:
Any third party estimation (paid or free) has the same problems since they are rely on the same approach: data from toolbar, from selected panel or from ISP. Some better because they have better coverage of the net population or geo regions.
Google has the best data since their toolbar is distributed more widely …
Only the raw traffic log of web sites give you 100% good data:-)
-
Adam Bruce said:
I can’t speak for others but Compete says we (StreetFire.net) only have 200K Unique monthly visitors, while our ad server and Google Analytics both say we have 3-4 Million. I’ve seen the server logs so I know there is no way 200K users create that much traffic to our site. Conversely Alexa shows us comparable to CarDomain.com who has equal traffic to us (We have a partnership in place with those guys so we know how our actual traffic ranks with theirs). I find it odd that Alexa would show us close to equal, but compete shows us totally not equal.
Anyhow here is what I would like to see, and I would pay money for this. Give out a service like Alexa for free, but then also give out tag-code like Google Analytics, and for those users that install the tag code on their site, they in turn get Google Analytics type views into their own traffic, plus other sites as well.
I would then be willing to pay money for deeper stat reporting onto our own site.
….or maybe google Analytics can just make some of their data public…….
-
Adam Bruce said:
UPDATE: Apparently I need to familiarize myself with these services a little more. Apparently quantcast.com does exactly what I described above…. of course they to are reporting us waay lower than we actually do so I think they must be pulling from the same data that Compete does.
-
TJ Mahony said:
Matt,
To expand on Lauren’s comments, it’s important that you note that Compete’s projections are ‘US projections’. Compete is a well established business that provides incredibly detailed competitive analysis to over 50 of the largest business in the U.S.. In addition, Compete’s numbers are actively used in the investment community to predict revenue of online centric businesses. In fact, Compete is a certified resource of Seeking Alpha and are publicly recognized by several financial institutions as a superior source of insight. I invite you to read some of the unsolicited public statments made on Compete’s behalf at: http://www.compete.com/testimonials
Thanks for supporting the conversation.
TJ
-
Adam Bruce said:
Update #2: Okay, I’ve installed the QuanCast.com code on all our StreetFire.net pages. (at approx 11:00 EST on Thursday Evening)
I think all of us would like to see a real life comparison to what Comepte and QuantCast report vs what in-house analytics report. So I’m willing to drop the veil and share our actual in-house results with all of you, as well as a before and after on installing the Quantcast Code
Compete.com says we have 239,723 Unique Monthly Visitors
http://snapshot.compete.com/streetfire.netQuantCast says we have a little over 100,000 monthly Uniques.
http://www.quantcast.com/streetfire.netAlexa says we have a reach per million of 180..
http://alexa.com/data/details/traffic_details?url=streetfire.netHere is a Screen shot of our Google Analytics Page showing 4.2 Million/month
http://assets.streetfire.net/traffic.png
http://assets.streetfire.net/traffic2.pngThis says we did 4.2 million Uniques Vidiac wide (across all our sites). Both our in house Ad Server and Analytics agree that roughly 71% of our Unique visitors go to StreetFire.net which is roughly 3.0 Million visitors plus or minus a few hundred thousand.
So I guess we can check back in a couple of days to see if Quantcast Changes their numbers following pasting their ad-code? All that said I would like for any one of these providers to justify the magnitudes of difference between what they’re reporting vs what we’re actually getting. NOTE: The Google number is at lest in the same BallPark +/- 20% of our server logs, which we correlate with Unique visitors, video streams, uploads, page views, User accounts, etc.
-
Otis Gospodnetic said:
Accuracy aside, there is one thing neither Compete nor Quantcast have, and Alexa does - timeliness! Alexa updates more or less daily, while the other two are over a month behind. I can look at Alexa and get a rough idea about the traffic (spike, drop, etc.) of a site from last week or less. With Compete and Quantcast one has to wait over a month. It is almost mid-February and Quantcast and Compete still show only December data!
-
Adam Bruce said:
Well, I have a comment waiting moderation right now due to the number of web links, bu I’m willing to throw down the gauntlet and show everyone waht or in-house reporting is saying vs, what these services are saying, and I would like to see what TJ Mahoney has to say as justification for the sher magnitudes of diference between the too. In a nutshell, they claim we’re doing 239K Uniques a MONTH, but our web stats show we do that in a DAY, and our Monthly is more like 3 Million Unique Users. So Hopefully Matt will approve my posts so you guys can see screen shots of our aalytics vs their reports.
-
TJ Mahony said:
Adam,
I’d be happy to take a look and provide an explaination. I assume the site you are referring to is streetfire.net
Feel free to email me (tmahony(at)compete.com)your graphs and data points. My first thought was that you attract a high concentration of international users. I took a quick look at Alexa (which is the only decent proxy for international traffic), as a point of comparision, and they measure your site 50% less than Compete.com. Therefore it seems that intl’ dynamics are not the cause, since Compete currently attracts ~200K UVs/month.
My gut tells me you are looking at visits or hits and not unique visitors; however, this is only a guess. Again, feel free to email me.
TJ
-
Andrew Holt said:
TJ, does Compete offer an API for its data? We would be very interested in talking to you about this.
-
Adam Bruce said:
TJ,
I will email you shortly with links to the snap shots, but those are our Uniques, not “hits”. (4.2 Million Uniques of which 71% went to StreetFire the remaining 29% went to other Vidiac.com affiliates), our page views for January for StreetFire were 45,840,994 page Views. StreetFire.net streamed (in-page not embed players) 24 Million videos in January. Actual “hits” were waaaaaaaaaay more than 45 Million last month. I’ll put it this way, our 95th percentile Hosting Bill was 1.4 GIG. So 230K users that Compete.com is reporting. is orders of magnitude off. On International 59.42% United States, 10% UK and Canada, So really our Compete number should be 1.79M, but the fact it’s reporting 239K? Emailing you screen shots now….-Adam
-
bring down alexa said:
I can not understand how alexa got any attention in the first place. It is so flawed and misguiding..if someone wants some proof we can send our own google analytics data and alexa rank.. another major problem- rank shown on Alexa site considers global visitors..it no way reflects any meaningful information ..there are quite a few sites whose rank is way high on alexa just because of global visitors.. For US based sites revenue model is almost always based on US visitors but alexa rank is based on Indian/chinese visitors.. btw, this is one of the reasons blog advertisements don’t yield expected results (at least that was our experience when we tried a few well known blogs with very high alexa rank..turned out most of the traffic was from “outsiders”)..there has to be some good service to give traffic analysis by state/city/demographics because that is what really matters ..who cares about global rank..may be bloggers can start the trend of making their google analytics data public..
-
TJ Mahony said:
Andrew:
Yes, we do offer an API. The first version of our API will go live next week. I am actually a registered member of Competitious, so I would be happy to talk about providing access. Just shoot me an email (see previous post for my address).
Best,
TJ -
Matt Marshall said:
Adam, I looked at my moderation records, and there’s nothing in the queue. Not sure what happened. You can try again, or you can send me the graphics, and I’ll post manually.
-
Matt said:
We’ve been very impressed with Quantcast. I’ve been testing their code across WordPress.com blogs and its numbers are more accurate than any other public measure I’ve seen and the demographic data it provides is really interesting.
I also like how they retroactively correct and re-rank sites based on new data that they get in. It’s in everyone’s interest to have the most accurate data out there as possible.
-
Search Engines WEB said:
seoptimization.blog.com/1221628/
here is a list of the top popular blogs listed by Technorati about 20 of them make their stats public. They represent a wide range of subjects
You can make a comparison with the estimates of Compete and Quantcast estimations to see which is the most accuarte
-
Julie Colwell said:
While Quantcast and Compete both claim to deliver demographics information, it’s estimated, not validated. And advertisers have to look up zillions of sites and compile that information themselves to take to publishers before they buy ads. There is only one company that can provide validated demographic information AND ad network partners, so advertisers can search for their target audience and make an ad buy at the same time. Internet research company IPRO. IPRO recently launched their advertising marketplace, SPOTSITE, where advertisers can shop for specific demographics across the thousands of researched web sites.
-
Pan_theFrog said:
While I am a little late to this topic, I’ve been watching certain sites that are quantified by Quantcast who seem to get next to no visitors, but Quantcast keeps moving them up in rank.
archimedes-lab.org is listed as having had one hit since Jan 23, but estimated to have 10k unique US visitors a month. Rank 138,983.
pimpnflyguy.com shows 5,147 US visitors with no more then 4 visitors a day. Rank 242,598 (Up 4k from yesterday)
betterworld.net shows 4,677 US visitors with no more then 25 visits a day. Rank 264,460.
Any of this making sense? -
qjanrpzrzn said:
Hello! Good Site! Thanks you! fmpgazclihjs
-
Mark said:
Compete underestimates us by two-thirds, but at least when our traffic goes up, their estimate goes up (and when traffic goes down, their estimate goes down).
Alexa has a mind of its own and doesn’t seem to represent reality at all, as represented by our logs and by Google Analytics and other log analysis programs.
Quantcast has us underestimated by about 60%, but the trend seems accurate.
Alexa is just broken, and I can’t believe Jeff Bezos is letting the company rot. They have thrown away their headstart. And that silly feud with Statsaholic: what a distraction, given the opportunity that they should be concentrating on.
-
Traffic Court said:
We have seen similar errors in Compete and Alexa for our traffic court website. Our traffic has gone up dramatically in the last month (per Google Analytics), yet the stats on Alexa and Compete show us as flat or declining.
-
Chris said:
My website went up minimally, however, now it is on a steady downward change, even though the number of alexa users has gone up, and the number of overall users has gone up. WTF? Any ideas?
-
mike said:
So its safe to say that Quantcast and compete and are saying a lesser amount of people are visiting any given site?
