More Stats on Del.icio.us, This Time Positive

Trackbacks/Pings (Trackback URL)

1. Austin
   

   August 4th, 2006 at 4:52 pm

   > . . . any perceived lack of new feature launches is due to a move to the Yahoo MyWeb platform from the legacy del.icio.us platform

   Even though many people claim that del.icio.us is ugly, it’s simple and usable. The MyWeb service, on the other hand is typically Yahoo — an attempt at oversimplification that generally ends up making the service less usable. Depending on the results of this “move,” I may have to find a new service.

2. rick gregory
   

   August 4th, 2006 at 5:17 pm

   Mike,

   Could the comscore data issue be partly explained by the fact that they’re tracking delicious.com and not del.icio.us? I imagine a LOT of links and organic traffic, as well as blog mentions, go to the latter domain.

3. Deepak
   

   August 4th, 2006 at 5:42 pm

   I still don’t understand the whole MyWeb thing. I have tried it multiple times and have never stayed. Yahoo needs to figure out which product they want to spend time on, and if I had to pick one, it would be del.icio.us

4. Danny
   

   August 4th, 2006 at 6:31 pm

   Hey Michael, to help us better understand Comscore’s methodology and to compare their stats to others, can you please post or comment on Comscore’s analysis for del.icio.us (not delicious.com)….

   It’s hard to judge potential inaccuracies without a reasonable comparison…

   Thanks,
   DK

5. Mike
   

   August 4th, 2006 at 6:49 pm

   I would go with the Yahoo! figures and not trust ComScore one bit. Back when I used to run New Ventures Strategy for eBay we ran into the same problems. ComScore showed that craigslist was not growing much but we believed otherwise and when we finally received the data from them for due dilligence we could see huge growth. ComScore is a sample based system and their methodologies and sample recruiting make them terrible for new web apps. So, I would not quote them.

   Overall, I am glad to to see that the category is growing. If anyone is looking for an alternative to del.icio.us, you might like to check out BlinkList.com

6. Markus
   

   August 4th, 2006 at 7:16 pm

   Comscore is completely worthless, they barely have an install base of 2x alexa. You quoted my USA traffic of 100 million pageviews a month for plentyoffish.com for the month of June using comscore data. For the month of May comscore said I had 200 million pageviews in the USA. Basically comscore says I lost 50% of my traffic in one month which is completely ridiculous.

   Comscore is considered a spyware application because It breaks HTTPS and reroutes all traffic through their servers. American banks block users who have the comscore spyware application installed from even accessing online banking. (Check article on MSNBC) In terms of accuracy comscore is barely above alexa, and doesn’t even come close to hitwise data.

7. Anshul
   

   August 4th, 2006 at 7:44 pm

   It is really surprising to see that stats differ so much between yahoo and other stat sites. What concerns me even more is the fact that though its ok to trust yahoo but isnt it very obvious that yahoo will always want to paint a nice picture of their own services.

8. Michael Arrington
   

   August 4th, 2006 at 7:59 pm

   Danny - the problem is that comscore doesn’t track icio.us, and it is unclear as to whether delicious.com is supposed to subsume that.

9. Michael Arrington
   

   August 4th, 2006 at 8:01 pm

   Markus,

   Yeah, but I can live with comscore not getting the number right. What worries me more is being inconsistent. So if they show 100m page views for you and then drop to 50m the next month, that’s worse than if they just showed 100m but at least reflected roughly your actual growth rate.

10. Clark
    

    August 4th, 2006 at 10:03 pm

    Yahoo figure will be trust worthy.

11. DrumsNWhistles
    

    August 4th, 2006 at 10:16 pm

    I don’t think any metrics service has a grip on how to handle sites like del.icio.us (or delicious.com) where so much of the traffic is RSS based. I add between 3-6 links per day to my del.icio.us with comments, which then feed on my main and sub RSS feeds.

    I have no idea how many people might click those links and view the articles I’ve bookmarked. I doubt that anyone does. At best, the site hosting the article might record entry via del.icio.us but that’s the extent of it.

    My first thought when reading your initial report was that the metrics didn’t account for RSS traffic. Until they wrap themselves around accurate measurements metrics are almost meaningless.

12. Markus
    

    August 4th, 2006 at 10:32 pm

    Micheal comscore showed my site as having 20 million/pageviews a month in the USA for the month of march, 100 million in april, 200 million in may and then 100 million in june. In actual fact my traffic has increased about 30% since march, slow steady growth in the off season to about 240 million US pageviews/month

    For my traffic 1 in 600 has alexa installed, 1 in 350 unique visitors has comscore installed. As for how they report Comscore reports absolute unique visitors per month, hitwise reports average dailly visitors. There have been several instances in the last few years where one page affiliate spam sites doing nothing but popup advertising showed up in the 10 most trafficed dating sites with “millions of uniques” a month on comscore.

    I think at the end of the day what this industry needs is some standards. There is no way to compare apples to apples between different providers. Some providers like nelson netratings and comscore compare apples to apples but their numbers often differ by a factor of 10.

13. Danny
    

    August 4th, 2006 at 10:37 pm

    Michael - that’s helpful, thx for the clarification, appreciate it…

14. Brian Mulvaney
    

    August 5th, 2006 at 12:02 am

    Comscore claims 2M+ panel members. These are people who installed the Marketscore monitoring application:

    “In exchange for having their Internet browsing and purchasing activity monitored, members have access to free email virus scanning and other benefits.”

    http://www.marketscore.com/

    Now does that sound like something del.icio.us users would install?

    Comscore’s sample size for del.icio.us is likely to be from an exceedingly small number of users. (PCs actually.) That could easily explain the big swings in Comscore projected traffic from period to period. Definitely worthless for this kind of use.

15. Otis
    

    August 5th, 2006 at 12:14 am

    Perhaps Comscore should be the next company to review here.

    A little more on traffic. Over at Simpy.com I split traffic generated by humans from that generated by aggregators and that generated by bots. This lets me look at these 3 traffic segments separately, and produces a chart like this. Note that the aggregator traffic is very close to the human traffic, while the bot traffic dwarfs them both. I also described the simple Apache config to split traffic like this here.
    Looking at bot and aggregator traffic alone is also interesting, as it exhibits very distinct nocturnal patterns, quite different from the nice little bell graphs one gets from human traffic, at least here in the U.S.
    I hope this helps others.

16. james kent
    

    August 5th, 2006 at 12:21 am

    More quality journalism from Michael Arrington.

    Brian is correct - these external monitoring sites do not necessarily represent a true cross section of online society.

    It would be very difficult for a publicly quoted service like Yahoo to lie about their statistics - if they were caught doing so, it could be fraud. I suppose they don’t have to release anything unless required to do so, but as a listed company they have to correct anything published about them that is materially incorrect, and that could affect the market. That could easily include hit stats, as that affects advertising, which is part of their core revenue….

    So perhaps this is a useful tactic to get companies to release their stats - just quote something spurious and make it a news story.

17. Shaun
    

    August 5th, 2006 at 1:56 am

    Going off Comscore and Alexa stats is pretty pointless IMHO.

    Firstly, del.icio.us is something the more web savvy user would use. I can’t imagine they’d also install Comscore or Alexa.

    Secondly, the alpha-geeks that use these services tend to feature Macs more heavily than the rest of the market and since Comscore and Alexa don’t count Macs, the stats are going to be meaningless.

    Thirdly, the trends are easily skewed by new browsers that don’t support the toolbar or Comscore/Alexa users getting tired of that toolbar and installing say the Google toolbar instead. As new toolbars are released, the share each toolbar has changes and that affects Comscore/Alexa stats.

    Hitwise on the other hand may be more representative since it doesn’t rely on toolbar installs although they’re still not accurate as they only track sites that are registered with Hitwise and they collect their data from more mainstream ISPs, again cutting out the alpha-geeks.

    I’ve pointed this out on a number of occasions to sites wearing the “We’re No.1″ Hitwise badge that they’re only No.1 because their competitors aren’t registered. It’s quite funny when the month after, they’ve dropped to No.3 because I’ve registered their competitors and their graphs tank. Clueless newbs.

18. Chris Lake
    

    August 5th, 2006 at 2:40 am

    Brian has hit the nail on the head - this is more about demographics than anything else.

    Alexa: useful for trends, not for absolutes. I run one tech blog and one entertainment blog. The latter attracts a far greater amount of daily unique users and page impressions, sometimes 20X the daily amount we’ll see on the tech blog. Alexa never reflects this - the ents blog will show spikes, but the tech blog is *always* ranked higher. Alexa is skewed towards tech. It ain’t accurate.

    Comscore: 2m installed users, plus survey-based data, and 2m is a lot of people. Should provide a good cross-section, but as Brian says, the people buying into this aren’t going to be particularly savvy if they are lured by ‘free email virus scanning’. So maybe we can say Comscore is skewed against tech. It too ain’t accurate.

    Hitwise: Watches something like 20m web users across various territories. That’s a mega-sample. However, the Hitwise data seems limited to consumer behaviour. As I understand it, data is sucked up from consumer ISPs, so at-work usage isn’t tracked. More useful for B2C companies, but not so good for B2B. Presumably no real view of lunchtime web activity, slackers-at-work, 9-5 search activity, etc. So it isn’t accurate either.

    Google & Yahoo: Do not release data on third party websites. But they both have widely-installed toolbars and as such are best placed to provide this sort of data. Either firm could provide stats that would be far more accurate than any of the above. I see Google as a huge advertising network, so sometime, sooner or later, I expect it to help advertisers figure out which sites to advertise on. Yahoo may follow suit. What’s the difference between somebody who installs a Google toolbar versus a Yahoo toolbar? Demographics will again come into play…

    In summary the tools available can be used in conjunction with one another, to improve accuracy (at a potentially high cost). But can any of these tools be considered truly ‘accurate’, when taking into account the big picture view?

    They’re good for trends, and best when compared against peer websites. But not for absolutes, in my view…

19. Jeremiah McNichols
    

    August 5th, 2006 at 4:49 am

    Michael,

    I’m really having a hard time understanding why you reported the Comscore data so vigorously. If I recall (haven’t looked back) you noted the delicious.com issue even then but it was more or less buried in the story. You did a good job keeping the communication channels open for a friendly correction from Yahoo!, but it seems obvious to me (maybe so obvioius I’m wrong!) that all traffic that is not sent there directly from some Yahoo! portal, which still includes a majority of users, goes to del.icio.us, not delicious.com, as do ALL links, feeds, etc. etc. etc. Rick Gregory points this out above but you already knew it. I think a heavier dose of skepticism would have been warranted from the gate, which might have made for a less sensational story. That said, I’m a big fan of your blog and am just a bit surprised to see you take this one for the team.

20. www
    

    August 5th, 2006 at 6:22 am

    Wow, interesting: Another example of the classic blog post “there is a rumor and I was unable to confirm the rumor, so let’s just keep spreading”. That’s what I call relevant news

21. James
    

    August 5th, 2006 at 7:18 am

    I use del.icio.us all the time and love it. Aside from any comments on statistics, my only complaint is that they still need to speed things up. Depending on the time of the day, there continues to be an annoying lag between the time you change a tag or bundle contents and the time it takes to actually show up correctly on the web. Hopefully Yahoo will put some money into further improvements, since things aren’t quite “there” yet.

22. Shareholder
    

    August 5th, 2006 at 8:42 am

    As a Yahoo shareholder, this makes the past quarter and the drop in the stock price seem much more palatable. I’m glad that for millions of dollars, the site has received 10% registration growth month over month and now has over 53 million posts (25 million URLS!!!!). I think Wall Street analysts are foolish for not looking at the value inherent in del.icio.us registrations and posts. And .us domains are poised to be the next hot market, so icio.us has some value too. Yahoo’s stock will rise again once the analysts realize this. We’re in a new economy and it’s not all about profits!

23. Saket Kumar
    

    August 5th, 2006 at 3:18 pm

    I may be one of the odd person who uses “My Web” more than del.icio.us, mainly because delicious didn’t had feature to make and post public, friends, private. But, after some time delicious brought all the features of My Web.

    So, one thing I’m still confused the why Yahoo want to compete with their own department and that’s also in public. They could have 2-3 department internally to compete with each other but let the better department win and have only one product in the market. Similarly, it’s happening in the case of Flickr and Yahoo photos.

24. Otis
    

    August 5th, 2006 at 4:53 pm

    Saket: they don’t compete with each other. They cover/attract mostly different segments of the population. Same with Flickr and Y! Photos.

25. Jeb
    

    August 5th, 2006 at 7:09 pm

    comScore is terrible.

    I run a site that got 1.1M visitors in June (per Google Analytics “absolute unique visitor” count), but comScore (I’m told by sources who can afford their reports) only showed us getting 750k visitors.

    I was telling a VC about the descrepancy and he was like, “yeah, comScore is way off because of their sampling method.”

    A 31% lower count??? That is not acceptable. It’s practically defamation.

26. Esp
    

    August 6th, 2006 at 2:27 am

    One of the main reasons why d.i.o has increased user registrations and increased usage, is because of the many spammers who are creating dozens of accounts and putting in craploads of links in to the system in order to get exposure to their system…

    I speak from experience. I have written and sold scripts that create multiple d.i.o accounts and post bookmarks to them…

    Cheers,
    Esp.

27. Brian Mulvaney
    

    August 6th, 2006 at 8:09 am

    Jeb: 31% lower projected unique user count may be plenty acceptable from panel based audience measurement. You can get almost as much variance in uniques by using different web analytics tools on your own site.

    The juice in Comscore is not in their estimate of the size of a given site’s audience, but in the behavioral data: how people got to the site; what keywords they were searching on; how and whether they bought something; how much they bought; why they left the site and where they went afterwards; what other comparable sites they also visited; etc. Think of it as clickstream analysis on your competitors’ websites. In some markets, those insights are worth a lot of money.

28. George
    

    August 6th, 2006 at 5:47 pm

    Interestingly, 10% month on month growth is almost exactly the trend shown in the hitwise chart.

    If the internal claim and one external source say the same thing, I have no trouble believing it.

29. Al Chumley
    

    August 7th, 2006 at 1:01 pm

    Thanks for looking into this one some more. Any update on the Google Talk stats?

30. Jeb
    

    August 10th, 2006 at 1:29 pm

    Brian: I have no problem accepting the value of the clickstream data, but I operate in a market where each visitor is worth minimally $0.35. So if comScore says we get 350k fewer visitors than we really do, then that’s $1.4M less traffic value per year. Multiply that by a valuation multiple and we’re talking serious money, at least for a bootstrapped enterprise.

    I think the problem is that bizdev and media people think that comScore is the final word on visitor counts, when it is not.

    I disagree with the assertion that different analytics trackers like Google Analytics vs. Webtrends Live vs. Coremetrics vs. WebSideStory would produce as much variance. Perhaps this is true if you’re using web server log data, but results from cookie-based tracking is what it is.

31. Minty
    

    November 14th, 2006 at 8:40 pm

    comScore screws the pooch again…

    “One point of controversy was around Digg’s claim of 20 million unique monthly visitors and steep monthly growth, whereas the Comscore’s most recent September report shows only 1.3 million monthly unique visitors and flat growth since April (see chart below). Comscore is notoriously flaky, and these numbers are for U.S. households only. Comscore is almost certainly significantly under-reporting Digg traffic.”

    http://www.webpronews.com/topn.....asure.html