Loading ...
Sorry, an error occurred while loading the content.
 

comScore Report

Expand Messages
  • Joe Wilson
    comScore released an analysis of cookie deletion behavior this morning that runs the risk of getting buried in all the DoubleClick /Google news. The press
    Message 1 of 11 , Apr 16, 2007
      comScore released an analysis of cookie deletion behavior this morning that runs the risk of getting buried in all the DoubleClick /Google news.

      The press release and summary findings can be found here - http://www.comscore.com/press/release.asp?press=1389

      The most interesting thing I found was that contrary to popular wisdom, 3rd party cookies do not seem to be deleted at a substantially higher rate than first party cookies.

      Oh, and Eric was quoted as well. ;o)

      Joe Wilson
      Chief Scientist
      TACODA
    • Judah Phillips
      Yeah, it s an interesting marketing press release. It does a good job positioning comScore against what it perceives to be it s competitors, and the messaging
      Message 2 of 11 , Apr 17, 2007
        Yeah, it's an interesting marketing press release. It does a good
        job positioning comScore against what it perceives to be it's competitors,
        and the messaging has received some traction.

        Things that make you go hhhhmmmmm:

        Participants in the comScore Global Network receive a package of benefits
        that have proven to be broadly appealing to all demographic segments. Among
        others, these benefits include:

        - Security software applications such as server-based virus
        protection, remote data storage, encrypted local storage, Internet history
        removal
        - Attractive sweepstakes prizes
        - Opportunity to impact and improve the Internet

        comScore gives it's panel "Internet history removal" software?

        Does that software remove cookies?


        [Non-text portions of this message have been removed]
      • Eric Peterson
        Joe, I had to read and re-read your comment about 3rd party cookies do not seem to be deleted at a substantially higher rate than first party cookies a few
        Message 3 of 11 , Apr 17, 2007
          Joe,

          I had to read and re-read your comment about "3rd party cookies do not
          seem to be deleted at a substantially higher rate than first party
          cookies" a few times ... interesting perspective!

          I think that while the comScore data does appear to level the playing
          field in terms of likelihood of deletion, I believe that third-party
          cookies are still being blocked at a higher rate than first-party
          cookies (blocked by P3P, etc.)

          Talk about compounding the problem, huh?

          Oh, I liked Curt's quote:

          "The comScore study confirms that relying on cookie counting alone for
          audience measurement can result in inflated unique visitor counts,"
          said Curt Viebranz, CEO of TACODA. "That's why TACODA is working with
          comScore to enhance and validate our cookie data with comScore's panel
          data."

          It is not hard to imagine a service from comScore or Quantcast that
          provides cookie deletion rates on an ongoing basis such that vendors
          or end-users can re-calibrate their unique visitor counts. Actually,
          forget I said that while I call my IP lawyer ;-)

          Eric T. Peterson
          Web Analytics Demystified






          --- In webanalytics@yahoogroups.com, Joe Wilson <joe.wilson@...> wrote:
          >
          > comScore released an analysis of cookie deletion behavior this
          morning that runs the risk of getting buried in all the DoubleClick
          /Google news.
          >
          > The press release and summary findings can be found here -
          http://www.comscore.com/press/release.asp?press=1389
          >
          > The most interesting thing I found was that contrary to popular
          wisdom, 3rd party cookies do not seem to be deleted at a substantially
          higher rate than first party cookies.
          >
          > Oh, and Eric was quoted as well. ;o)
          >
          > Joe Wilson
          > Chief Scientist
          > TACODA
          >
        • Joe Wilson
          Eric, I agree that 3rd party cookies continue to be blocked at a higher rate than 1st party cookies (7-10% vs 1-2% in our analysis). I see this as less of a
          Message 4 of 11 , Apr 17, 2007
            Eric,

            I agree that 3rd party cookies continue to be blocked at a higher rate than 1st party cookies (7-10% vs 1-2% in our analysis). I see this as less of a problem because that is a situation that is not difficult to detect and quantify. It doesn't necessarily lead to more accurate data, but it does lead to more consistent data.

            Cookie deletion is not detectable - at least not using acceptable methods - and so is difficult to quantify on a per site or network basis.

            I actually see this as a marginally encouraging result. That most consumers do not readily distinguish between 1st party and 3rd party cookies would tend to support the notion that consumers are not making an informed choice when deleting cookies, but believe that all cookies are "bad". I am not saying that the comScore data conclusively shows this (far from it), only that this is one way to interpret the data.

            I think Curt's quote simply echos what many folks on this forum and elsewhere have been saying for some time - in an environment that has a significant amount of inherent uncertainty, you have use multiple sources to triangulate on a result rather than relying on any one method exclusively.

            Regards,

            Joe Wilson
            Chief Scientist
            TACODA

            ----- Original Message -----
            From: Eric Peterson <eric.peterson@...>
            To: webanalytics@yahoogroups.com
            Sent: Tuesday, April 17, 2007 11:25:44 AM
            Subject: [webanalytics] Re: comScore Report

            Joe,

            I had to read and re-read your comment about "3rd party cookies do not
            seem to be deleted at a substantially higher rate than first party
            cookies" a few times ... interesting perspective!

            I think that while the comScore data does appear to level the playing
            field in terms of likelihood of deletion, I believe that third-party
            cookies are still being blocked at a higher rate than first-party
            cookies (blocked by P3P, etc.)

            Talk about compounding the problem, huh?

            Oh, I liked Curt's quote:

            "The comScore study confirms that relying on cookie counting alone for
            audience measurement can result in inflated unique visitor counts,"
            said Curt Viebranz, CEO of TACODA. "That's why TACODA is working with
            comScore to enhance and validate our cookie data with comScore's panel
            data."

            It is not hard to imagine a service from comScore or Quantcast that
            provides cookie deletion rates on an ongoing basis such that vendors
            or end-users can re-calibrate their unique visitor counts. Actually,
            forget I said that while I call my IP lawyer ;-)

            Eric T. Peterson
            Web Analytics Demystified






            --- In webanalytics@yahoogroups.com, Joe Wilson <joe.wilson@...> wrote:
            >
            > comScore released an analysis of cookie deletion behavior this
            morning that runs the risk of getting buried in all the DoubleClick
            /Google news.
            >
            > The press release and summary findings can be found here -
            http://www.comscore.com/press/release.asp?press=1389
            >
            > The most interesting thing I found was that contrary to popular
            wisdom, 3rd party cookies do not seem to be deleted at a substantially
            higher rate than first party cookies.
            >
            > Oh, and Eric was quoted as well. ;o)
            >
            > Joe Wilson
            > Chief Scientist
            > TACODA
            >




            ---------------------------------------
            The Web Analytics Forum
            Founded by Eric T. Peterson (www.webanalyticsdemystified.com)
            Moderated by the Web Analytics Association (www.webanalyticsassociation.org)
            Email moderators at: webanalytics-moderators@yahoogroups.com
            Yahoo! Groups Links
          • christopherjones16
            I find it interesting that the industry tries to come up with a % of overstatement by which we can all adjust our unique users to find out our true reach.
            Message 5 of 11 , Apr 17, 2007
              I find it interesting that the industry tries to come up with a % of
              overstatement by which we can all 'adjust' our unique users to find
              out our true reach. The affect on your sites unique user number is
              really more a function of loyalty over the given period.

              For arguement sake, let's assume that 10% of users delete all cookies
              1/month, another 10% delete their cookies 1/week, and another 10%
              delete their cookies 1/day

              Site 1: News Site, one that most users visit every day (M-F only) all
              year long.

              Site 2: Ecommerce site, one that users will visit several times/week
              but only for that 1 week period.

              Site 3: Niche site, one that users will visit occasionally here and
              there, but most likely only 1/month

              Site Monthly Uniques Monthly Uniques % Overstatement
              1 1000 250%
              2 1000 50%
              3 1000 0%

              I was asked by a previous employer to come up with are own
              internal 'adjustment' factors based on loyalty. The range across
              sites was almost as dramatic as the example above.

              Chris
              --- In webanalytics@yahoogroups.com, "Eric Peterson"
              <eric.peterson@...> wrote:
              >
              > Joe,
              >
              > I had to read and re-read your comment about "3rd party cookies do
              not
              > seem to be deleted at a substantially higher rate than first party
              > cookies" a few times ... interesting perspective!
              >
              > I think that while the comScore data does appear to level the
              playing
              > field in terms of likelihood of deletion, I believe that third-party
              > cookies are still being blocked at a higher rate than first-party
              > cookies (blocked by P3P, etc.)
              >
              > Talk about compounding the problem, huh?
              >
              > Oh, I liked Curt's quote:
              >
              > "The comScore study confirms that relying on cookie counting alone
              for
              > audience measurement can result in inflated unique visitor counts,"
              > said Curt Viebranz, CEO of TACODA. "That's why TACODA is working
              with
              > comScore to enhance and validate our cookie data with comScore's
              panel
              > data."
              >
              > It is not hard to imagine a service from comScore or Quantcast that
              > provides cookie deletion rates on an ongoing basis such that vendors
              > or end-users can re-calibrate their unique visitor counts.
              Actually,
              > forget I said that while I call my IP lawyer ;-)
              >
              > Eric T. Peterson
              > Web Analytics Demystified
              >
              >
              >
              >
              >
              >
              > --- In webanalytics@yahoogroups.com, Joe Wilson <joe.wilson@> wrote:
              > >
              > > comScore released an analysis of cookie deletion behavior this
              > morning that runs the risk of getting buried in all the DoubleClick
              > /Google news.
              > >
              > > The press release and summary findings can be found here -
              > http://www.comscore.com/press/release.asp?press=1389
              > >
              > > The most interesting thing I found was that contrary to popular
              > wisdom, 3rd party cookies do not seem to be deleted at a
              substantially
              > higher rate than first party cookies.
              > >
              > > Oh, and Eric was quoted as well. ;o)
              > >
              > > Joe Wilson
              > > Chief Scientist
              > > TACODA
              > >
              >
            • Eric Peterson
              Hi Chris, I absolutely agree with your assessment of the problem but I m not sure the industry is trying to come up with a percent overstatement that can be
              Message 6 of 11 , Apr 18, 2007
                Hi Chris,

                I absolutely agree with your assessment of the problem but I'm not
                sure the "industry" is trying to come up with a percent overstatement
                that can be used to correct for the cookie deletion effect.
                Especially given that the primary research addressing cookie deletion
                has come from Nielsen and comScore (both of whom have a vested
                interest), I think that we're still very much at the "communicate that
                the problem exists and roughly scope the magnitude" phase.

                Yes, I think this is ironic (or perhaps just unfortunate) given that
                this research has been around for a pretty long time (the RedEye
                Report and others pre-dated my work at JupiterResearch, etc.)

                While we in this group pretty much seem to agree that the problem
                exists, I have to wonder how widely the message is communicated
                outside of the "analytics enclave" I mean, who of us when faced with
                presenting to the CEO or Board of Directors actually says "We had
                1,000,000 unique visitors in April, but because of cookie deletion and
                the serial deleter effect, the real number of people who came to the
                site could be a lot closer to 400,000, or maybe 600,000, but it's
                probably really not a million, sorry ..."

                Now it sounds like you have a pretty good handle on the situation, but
                we all know that for the most part we just don't know how many actual
                people are visiting our sites.

                And this wouldn't be a problem, but so often when we talk in terms of
                "unique visitors" what is being conveyed is "people" not "cookies" or
                "browsers" or some technical abstraction that ** we in this forum **
                are pretty comfortable with but that doesn't always translate well
                outside of web analytics.

                I've said it before and I'll say it again: The technology firm that is
                willing and able to publish monthly cookie deletion numbers by
                industry vertical and/or sub-vertical AND share their methodology for
                coming up with the numbers for critique by the entire community will,
                if nothing else, deserve our heartfelt thanks. I suspect they could
                actually sell the data if they wanted to and thusly earn our
                hard-earned money as well.

                I know of at least one start-up that has this capability, there are
                very likely more.

                Cheers,

                Eric T. Peterson
                Web Analytics Demystified




                --- In webanalytics@yahoogroups.com, "christopherjones16"
                <christopher_jones@...> wrote:
                >
                > I find it interesting that the industry tries to come up with a % of
                > overstatement by which we can all 'adjust' our unique users to find
                > out our true reach. The affect on your sites unique user number is
                > really more a function of loyalty over the given period.
                >
                > For arguement sake, let's assume that 10% of users delete all cookies
                > 1/month, another 10% delete their cookies 1/week, and another 10%
                > delete their cookies 1/day
                >
                > Site 1: News Site, one that most users visit every day (M-F only) all
                > year long.
                >
                > Site 2: Ecommerce site, one that users will visit several times/week
                > but only for that 1 week period.
                >
                > Site 3: Niche site, one that users will visit occasionally here and
                > there, but most likely only 1/month
                >
                > Site Monthly Uniques Monthly Uniques % Overstatement
                > 1 1000 250%
                > 2 1000 50%
                > 3 1000 0%
                >
                > I was asked by a previous employer to come up with are own
                > internal 'adjustment' factors based on loyalty. The range across
                > sites was almost as dramatic as the example above.
                >
                > Chris
                > --- In webanalytics@yahoogroups.com, "Eric Peterson"
                > <eric.peterson@> wrote:
                > >
                > > Joe,
                > >
                > > I had to read and re-read your comment about "3rd party cookies do
                > not
                > > seem to be deleted at a substantially higher rate than first party
                > > cookies" a few times ... interesting perspective!
                > >
                > > I think that while the comScore data does appear to level the
                > playing
                > > field in terms of likelihood of deletion, I believe that third-party
                > > cookies are still being blocked at a higher rate than first-party
                > > cookies (blocked by P3P, etc.)
                > >
                > > Talk about compounding the problem, huh?
                > >
                > > Oh, I liked Curt's quote:
                > >
                > > "The comScore study confirms that relying on cookie counting alone
                > for
                > > audience measurement can result in inflated unique visitor counts,"
                > > said Curt Viebranz, CEO of TACODA. "That's why TACODA is working
                > with
                > > comScore to enhance and validate our cookie data with comScore's
                > panel
                > > data."
                > >
                > > It is not hard to imagine a service from comScore or Quantcast that
                > > provides cookie deletion rates on an ongoing basis such that vendors
                > > or end-users can re-calibrate their unique visitor counts.
                > Actually,
                > > forget I said that while I call my IP lawyer ;-)
                > >
                > > Eric T. Peterson
                > > Web Analytics Demystified
                > >
                > >
                > >
                > >
                > >
                > >
                > > --- In webanalytics@yahoogroups.com, Joe Wilson <joe.wilson@> wrote:
                > > >
                > > > comScore released an analysis of cookie deletion behavior this
                > > morning that runs the risk of getting buried in all the DoubleClick
                > > /Google news.
                > > >
                > > > The press release and summary findings can be found here -
                > > http://www.comscore.com/press/release.asp?press=1389
                > > >
                > > > The most interesting thing I found was that contrary to popular
                > > wisdom, 3rd party cookies do not seem to be deleted at a
                > substantially
                > > higher rate than first party cookies.
                > > >
                > > > Oh, and Eric was quoted as well. ;o)
                > > >
                > > > Joe Wilson
                > > > Chief Scientist
                > > > TACODA
                > > >
                > >
                >
              • Jackson, Steve
                Hi all, Been following this with interest and agree with most of the sentiments here. One comment was made (by Eric I think) which said what if your business
                Message 7 of 11 , Apr 19, 2007
                  Hi all,

                  Been following this with interest and agree with most of the sentiments
                  here. One comment was made (by Eric I think) which said "what if your
                  business model depends on unique visitors?". This is the case with many
                  of my clients who run media websites for instance. However in their case
                  the solution is not to do true web analytics but rather to rely on one
                  tool (as Anil said). For instance the IAB in finland (and another
                  organisation in the Nordics) basically track all the media websites via
                  TNS gallup's tool.

                  http://www.gallupweb.com/tnsmetrix/

                  So every week we get cookie corrected data based on a standard
                  correction method that TNS use for all sites and the figures taken from
                  TNS metrix ASP system which all the media sites use. (TNS metrix is a
                  low end web analytics system that tracks the basics as well as any other
                  ASP).

                  Now this used to bug the hell out of me to be honest because TNS were
                  initially in my opinion self serving and blocking the media portals from
                  using web analytics so that they wouldn't lose business (as in reality
                  each portal would need to invest in two tools, one to stay on the media
                  list and another to get a good tool). However I've changed my opinion in
                  part because we needed to educate the media portals about the benefits
                  of web analytics rather than bash TNS for not having a good web
                  analytics solution. This is why 18 months of WAA activities in the
                  Nordics have started to pay off and now the market (at least in Finland)
                  is waking up to the difference between audience measurement and web
                  analytics.

                  It means the businesses who rely purely on advertising income can use
                  TNS Metrix's published list with confidence that they are being compared
                  fairly across the media space and the advertisers who pay the media
                  portals have the same level of confidence. The UV figures whether they
                  are 100% correct or not become largely irrelevant.

                  Those that also need in depth analytics look elsewhere for other tools
                  (thanks to WAA education). If the rest of the world got the IAB or
                  similar organisations to set-up and run the same kind of lists then the
                  problem would be effectively solved.

                  Br
                  Steve Jackson

                  Senior Consultant

                  Web Analytics & Search Marketing



                  SATAMA

                  Henry Fordin katu 6

                  FI-00150 Helsinki Finland

                  tel. +358-207-581 858

                  fax. +358-207-581 893

                  mobile +358-50-343 5159

                  email steve.jackson@...

                  www.satama.com <http://www.satama.com/>

                  www.aboavista.com <http://www.aboavista.com/>

                  www.conversionchronicles.com <http://www.conversionchronicles.com/>



                  Nordic Co-Chair

                  Web Analytics Association

                  http://www.webanalyticsassociation.com
                  <http://www.webanalyticsassociation.com/>



                  LinkedIn

                  www.linkedin.com/in/webanalytics007





                  ________________________________

                  From: webanalytics@yahoogroups.com [mailto:webanalytics@yahoogroups.com]
                  On Behalf Of Eric Peterson
                  Sent: 18. huhtikuuta 2007 15:32
                  To: webanalytics@yahoogroups.com
                  Subject: [webanalytics] Re: comScore Report



                  Hi Chris,

                  I absolutely agree with your assessment of the problem but I'm not
                  sure the "industry" is trying to come up with a percent overstatement
                  that can be used to correct for the cookie deletion effect.
                  Especially given that the primary research addressing cookie deletion
                  has come from Nielsen and comScore (both of whom have a vested
                  interest), I think that we're still very much at the "communicate that
                  the problem exists and roughly scope the magnitude" phase.

                  Yes, I think this is ironic (or perhaps just unfortunate) given that
                  this research has been around for a pretty long time (the RedEye
                  Report and others pre-dated my work at JupiterResearch, etc.)

                  While we in this group pretty much seem to agree that the problem
                  exists, I have to wonder how widely the message is communicated
                  outside of the "analytics enclave" I mean, who of us when faced with
                  presenting to the CEO or Board of Directors actually says "We had
                  1,000,000 unique visitors in April, but because of cookie deletion and
                  the serial deleter effect, the real number of people who came to the
                  site could be a lot closer to 400,000, or maybe 600,000, but it's
                  probably really not a million, sorry ..."

                  Now it sounds like you have a pretty good handle on the situation, but
                  we all know that for the most part we just don't know how many actual
                  people are visiting our sites.

                  And this wouldn't be a problem, but so often when we talk in terms of
                  "unique visitors" what is being conveyed is "people" not "cookies" or
                  "browsers" or some technical abstraction that ** we in this forum **
                  are pretty comfortable with but that doesn't always translate well
                  outside of web analytics.

                  I've said it before and I'll say it again: The technology firm that is
                  willing and able to publish monthly cookie deletion numbers by
                  industry vertical and/or sub-vertical AND share their methodology for
                  coming up with the numbers for critique by the entire community will,
                  if nothing else, deserve our heartfelt thanks. I suspect they could
                  actually sell the data if they wanted to and thusly earn our
                  hard-earned money as well.

                  I know of at least one start-up that has this capability, there are
                  very likely more.

                  Cheers,

                  Eric T. Peterson
                  Web Analytics Demystified

                  --- In webanalytics@yahoogroups.com
                  <mailto:webanalytics%40yahoogroups.com> , "christopherjones16"
                  <christopher_jones@...> wrote:
                  >
                  > I find it interesting that the industry tries to come up with a % of
                  > overstatement by which we can all 'adjust' our unique users to find
                  > out our true reach. The affect on your sites unique user number is
                  > really more a function of loyalty over the given period.
                  >
                  > For arguement sake, let's assume that 10% of users delete all cookies
                  > 1/month, another 10% delete their cookies 1/week, and another 10%
                  > delete their cookies 1/day
                  >
                  > Site 1: News Site, one that most users visit every day (M-F only) all
                  > year long.
                  >
                  > Site 2: Ecommerce site, one that users will visit several times/week
                  > but only for that 1 week period.
                  >
                  > Site 3: Niche site, one that users will visit occasionally here and
                  > there, but most likely only 1/month
                  >
                  > Site Monthly Uniques Monthly Uniques % Overstatement
                  > 1 1000 250%
                  > 2 1000 50%
                  > 3 1000 0%
                  >
                  > I was asked by a previous employer to come up with are own
                  > internal 'adjustment' factors based on loyalty. The range across
                  > sites was almost as dramatic as the example above.
                  >
                  > Chris
                  > --- In webanalytics@yahoogroups.com
                  <mailto:webanalytics%40yahoogroups.com> , "Eric Peterson"
                  > <eric.peterson@> wrote:
                  > >
                  > > Joe,
                  > >
                  > > I had to read and re-read your comment about "3rd party cookies do
                  > not
                  > > seem to be deleted at a substantially higher rate than first party
                  > > cookies" a few times ... interesting perspective!
                  > >
                  > > I think that while the comScore data does appear to level the
                  > playing
                  > > field in terms of likelihood of deletion, I believe that third-party
                  > > cookies are still being blocked at a higher rate than first-party
                  > > cookies (blocked by P3P, etc.)
                  > >
                  > > Talk about compounding the problem, huh?
                  > >
                  > > Oh, I liked Curt's quote:
                  > >
                  > > "The comScore study confirms that relying on cookie counting alone
                  > for
                  > > audience measurement can result in inflated unique visitor counts,"
                  > > said Curt Viebranz, CEO of TACODA. "That's why TACODA is working
                  > with
                  > > comScore to enhance and validate our cookie data with comScore's
                  > panel
                  > > data."
                  > >
                  > > It is not hard to imagine a service from comScore or Quantcast that
                  > > provides cookie deletion rates on an ongoing basis such that vendors
                  > > or end-users can re-calibrate their unique visitor counts.
                  > Actually,
                  > > forget I said that while I call my IP lawyer ;-)
                  > >
                  > > Eric T. Peterson
                  > > Web Analytics Demystified
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > > --- In webanalytics@yahoogroups.com
                  <mailto:webanalytics%40yahoogroups.com> , Joe Wilson <joe.wilson@>
                  wrote:
                  > > >
                  > > > comScore released an analysis of cookie deletion behavior this
                  > > morning that runs the risk of getting buried in all the DoubleClick
                  > > /Google news.
                  > > >
                  > > > The press release and summary findings can be found here -
                  > > http://www.comscore.com/press/release.asp?press=1389
                  <http://www.comscore.com/press/release.asp?press=1389>
                  > > >
                  > > > The most interesting thing I found was that contrary to popular
                  > > wisdom, 3rd party cookies do not seem to be deleted at a
                  > substantially
                  > > higher rate than first party cookies.
                  > > >
                  > > > Oh, and Eric was quoted as well. ;o)
                  > > >
                  > > > Joe Wilson
                  > > > Chief Scientist
                  > > > TACODA
                  > > >
                  > >
                  >







                  [Non-text portions of this message have been removed]
                • Steve
                  ... ... ie. The problem is not cookie deletion rates or accuracy in a particular number. The problem is being fairly compared with ones competitors.
                  Message 8 of 11 , Apr 19, 2007
                    On 4/19/07, Jackson, Steve <steve.jackson@...> wrote:
                    > here. One comment was made (by Eric I think) which said "what if your
                    > business model depends on unique visitors?". This is the case with many

                    <snip>

                    > It means the businesses who rely purely on advertising income can use
                    > TNS Metrix's published list with confidence that they are being compared
                    > fairly across the media space and the advertisers who pay the media
                    > portals have the same level of confidence. The UV figures whether they
                    > are 100% correct or not become largely irrelevant.


                    ie. The problem is not cookie deletion rates or accuracy in a particular number.
                    The problem is being fairly compared with ones competitors. Which is a
                    very different issue, and realistically, irrelevant to cookie deletion
                    rates. Or page numbers or hits or any such metric.

                    ie2: It's a political problem, not a technical one.


                    The other half (Third? Quarter? Tenth? :-) ) of the problem in this
                    particular instance/market is (apparently) using a metric that means
                    one thing to imply something else. Using "Unique Visitors" to mean
                    "People". Unique Visitors is an *approximation* of People, but it is
                    NOT equal.[1]
                    Evil Grin. Unique Visitors == Unique Visitors. To belabour the point. :-)


                    A large danger is that by using a metric incorrectly [2] is that you
                    open it to abuse by the unscrupulous or desperate. In the same manner
                    that Hits can be inflated by more images; Pages inflated by splitting
                    articles across... err... pages :-) and so on. Unique Visitors can
                    also be manipulated upwards *if you really want to*.
                    Totally pointless to do so for oneself alone....


                    Again, to my mind, this just reinforces that in the cases you describe
                    Steve, the problem is a political one not technical. And in your 2nd
                    para I've left above, you describe the political solution. Most
                    succinctly in your last sentence, quoted above.

                    Accuracy doesn't matter - fairness does.


                    I'm completely ignoring the issue of "trust" in this latest raving.
                    But that also plays a key component. :-)


                    Cheers!

                    - Steve

                    [1] Pretty darn good approximation IMHO, but still approximation.
                    [2] Ignoring all the ramifications of same; Methodology errors; Market
                    shifts away from prior "accurate" methodologies; Different Markets
                    with different "rules" etc etc etc.
                  • Fulton Yancy
                    Steve, I wonder what the magical cookie corrected data based on a standard correction method really is and how it manipulates the collected metrics? Also
                    Message 9 of 11 , Apr 20, 2007
                      Steve,
                      I wonder what the magical cookie corrected data based on a standard
                      correction method really is and how it manipulates the collected
                      metrics?

                      Also please elaborate on why there even exists a need to manipulate
                      the collected metrics and what the benefits are. It really reflects
                      the mind set of Swedish public metrics when only 1 vendor was used,
                      if the system counted incorrectly everyone had the same variation of
                      correctness.

                      Rgds
                      Fulton

                      * I am a WebSideStory employee
                      * The above comments are my personal opinions
                    • Steve
                      Hey Fulton, No Worries. I d have thought it obvious to both questions... shows how wrong (again) I can be. :-) (b) Why? Fraud. Pure and simple. (a) How?
                      Message 10 of 11 , Apr 20, 2007
                        Hey Fulton, No Worries.
                        I'd have thought it obvious to both questions... shows how wrong
                        (again) I can be. :-)

                        (b) Why? Fraud. Pure and simple.
                        (a) How? Botnets for one. Plenty of similar methods around.

                        If you wanted to commit such fraud more directly, you would simply
                        look at how collection etc occurs and target their weak points.
                        Analytics is based around a level of trust in what you're getting
                        back. When it's purely for internal use there's not much point
                        manipulating the numbers like this. When you can manipulate one's
                        market share, the desperate may do stupid (and probably illegal)
                        things.

                        If the methodology used to determine share is "fragile" it can be
                        abused. And cookies are inherently fragile as they're stored client
                        side. ie Untrusted.

                        eg. Look at how "unique cookies" are determined. Rather than treating
                        them as static put a built in roller at the back end server side, that
                        gradually changes the coded [1] value on each and every visit. Or
                        faster. You haven't changed the actual # of people, but you've
                        manipulated the data collection - which is the target in this example.

                        I can expand further in tedious detail around a basic risk analysis
                        if you really want. But it's Saturday morning here, and I'd rather
                        not. ;-)


                        It's just a form of click fraud targeting "Market Share" vs adverts.
                        Same same. Nothing magical about it.


                        Hope that clarifies?

                        Cheers!
                        - Steve

                        [1] black side vs red if anyone understands crypto terms

                        On 4/20/07, Fulton Yancy <webbanalys@...> wrote:
                        > Steve,
                        > I wonder what the magical cookie corrected data based on a standard
                        > correction method really is and how it manipulates the collected
                        > metrics?
                        >
                        > Also please elaborate on why there even exists a need to manipulate
                        > the collected metrics and what the benefits are. It really reflects
                        > the mind set of Swedish public metrics when only 1 vendor was used,
                        > if the system counted incorrectly everyone had the same variation of
                        > correctness.
                        >
                        > Rgds
                        > Fulton
                        >
                        > * I am a WebSideStory employee
                        > * The above comments are my personal opinions
                        >
                        >
                        > ---------------------------------------
                        > The Web Analytics Forum
                        > Founded by Eric T. Peterson (www.webanalyticsdemystified.com)
                        > Moderated by the Web Analytics Association (www.webanalyticsassociation.org)
                        > Email moderators at: webanalytics-moderators@yahoogroups.com
                        > Yahoo! Groups Links
                        >
                        >
                        >
                        >
                      • Jackson, Steve
                        Ismo Tenkanen (TNS Gallup) published a white paper discussing how they correct the data. There is no need to manipulate data, but based on the studies I ve
                        Message 11 of 11 , Apr 23, 2007
                          Ismo Tenkanen (TNS Gallup) published a white paper discussing how they
                          correct the data.

                          There is no need to manipulate data, but based on the studies I've seen
                          they report more accuracy
                          at least when we talk about figures reported in Finland. For instance
                          you don't get weekly figures
                          with more people actually going to a website than exist in the country.

                          As I say though, it's not as relevant what the UV figures are it's more
                          a question of fair comparison.

                          We've had similar discussions before Fulton and we both know that this
                          is me changing my opinion.
                          I used to be strongly opposed to the monopoly like situation that
                          existed in finland with TNS. However
                          TNS fairly measure the audience numbers and we don't get people worrying
                          about cookies as a result.

                          Bizarre as it sounds TNS are helping the industry. I have never had to
                          fight to pitch any tool to anyone in
                          the last 12 months because of cookies. Simply put the TNS and TOY lists
                          in Finland have helped more
                          than hindered the web analytics industry in Finland.



                          ________________________________

                          From: webanalytics@yahoogroups.com [mailto:webanalytics@yahoogroups.com]
                          On Behalf Of Fulton Yancy
                          Sent: 20. huhtikuuta 2007 15:15
                          To: webanalytics@yahoogroups.com
                          Subject: [webanalytics] Re: comScore Report



                          Steve,
                          I wonder what the magical cookie corrected data based on a standard
                          correction method really is and how it manipulates the collected
                          metrics?

                          Also please elaborate on why there even exists a need to manipulate
                          the collected metrics and what the benefits are. It really reflects
                          the mind set of Swedish public metrics when only 1 vendor was used,
                          if the system counted incorrectly everyone had the same variation of
                          correctness.

                          Rgds
                          Fulton

                          * I am a WebSideStory employee
                          * The above comments are my personal opinions






                          [Non-text portions of this message have been removed]
                        Your message has been successfully submitted and would be delivered to recipients shortly.