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

12217Re: [webanalytics] Re: Valuing Online Video Properties with Web Metrics

Expand Messages
  • dhs1986@comcast.net
    Aug 1, 2007
      Lauren,

      Your methodology does not capture Digg's RSS audience, which is what the Digg CEO (rightly IMO) blamed in November 2006 for the relatively small unique visitor figures reported by the panel services. However, this does nothing to explain the disparity between the May Compete numbers and the panel services'.

      While by now everyone hopefully appreciates the difficulties involved in measuring people on the internet and your company's efforts to introduce new methodologies to the industry, I stand by my previous comments given the below:

      May 2007 U.S. Unique Visitors

      NetRatings NetView ComScore Compete
      Digg 4.3m 3.3m 23m
      Facebook 14.2m 26.6m 20m


      Best,
      David

      -------------- Original message --------------
      From: "Lauren Moores" <lmoores@...>

      In regards to David's post:

      Also there is some controversy with regard to the quality of Compete data
      due to possible concerns about sample bias (part of the data they collect is
      captured using a toolbar that anyone can install on their desktops). For
      example they recently reported on their blog with some fanfare that Digg
      attracted more visitors than Facebook in May, something that not only sounds
      highly implausible but is contradicted by every other source I checked
      (ComScore, NetRatings, etc), most of which report that Facebook drew more
      than twice as many visitors as Digg during that month.

      As troublesome as ComScore and Nielsen are, it's really starting to look as
      though the free traffic tools like Alexa, Compete, and Quantcast are worse
      IMO.

      For those of you unfamiliar with Compete prior to compete.com launch in
      November 2006, we have been providing granular consumer behavior segment
      analyses based on our 2M consumer panel for over six years to Fortune 500
      brands. We have a dynamic and distinctive methodology which takes data from
      many types of sources to create the best possible sample for projecting to
      the IBP (internet browser population). Our sources include Compete data
      collection through our recruited panels, our toolbar and our partnerships
      with clients. In addition, we integrate ASP and ISP data. The different
      sources that we use allow us to put our data in context and manage source
      bias better than a one panel approach.

      For the May numbers on Digg and Facebook, our Digg numbers are consistent
      with numbers reported by the Digg CEO.

      See
      http://www.techcrunch.com/2006/11/26/forbes-video-interview-with-digg-ceo/.

      As for Facebook, Compete tracks with Nielsen and comScore metrics and is
      spot on with Quantcast which is using their pixel technology with the
      Facebook site to track visitors.

      It would be a shame if people did not use our data based on a
      misunderstanding of our methodology and process. However, if you want to
      discuss this further, please feel free to email me separately.

      Thanks!

      Lauren

      On 7/31/07, ejt11three <ejthomas113@...> wrote:
      >
      > Thanks for the feedback, Ravi. I'm aware of the concept of a premium
      > paid to a market leader and the goodwill that accompanies. I mention
      > this in the fifth paragraph of my posting as follows:
      >
      > "In this example I use the YouTube's acquisition price and Compete's
      > April numbers to set a baseline. Keep in mind that YouTube's traffic
      > has grown significantly since its purchase by Google in November of
      > 2006, so this may lead to conservative cross-analysis. Nonetheless,
      > since YouTube is number one, and number deserves a premium, I assume
      > these effects will cancel each other out."
      >
      > This valuation is based on April traffic numbers, which I feel
      > discounts the November acquisition price. It's not a perfect discount
      > premium but I thought it would suffice.
      >
      > --- In webanalytics@yahoogroups.com <webanalytics%40yahoogroups.com>,
      > "Ravi Pathak" <ravipathak1@...>
      > wrote:
      > >
      > > Nice work !! The bigger assumption is that the rate at which google
      > paid for
      > > market leader (read youtube.com), other buyers are willing to buy at
      > same
      > > rate for followers.
      > >
      > > I dont think this holds true. Market leaders will always get premium and
      > > followers are generally laggards in terms of valuations.
      > >
      > > still not sure how much you will discount though.
      > >
      > > -R
      > >
      > > On 7/30/07, ejt11three <ejthomas113@...> wrote:
      > > >
      > > > There's been a lot of discussion of how to value hot online video
      > > > properties, like Youtube. There of course are many factors to this,
      > > > but I propose a simple valuation model by using common web metrics to
      > > > gauge user engagement. In other words, a simple way to value an online
      > > > video property is in terms of not only how big its audience is but
      > > > also how engaged they are (similar to a TV show). The more attention
      > > > we have, the higher the prospects are for gaining advertisement
      > > > revenue, and therefore the more valuable the property is.
      > > >
      > > > Using this model we can also gauge what other online video sites might
      > > > be worth by comparing user engagement relative to that of YouTube.
      > > >
      > > > We will use the common web metrics (i.e. visits and average time spent
      > > > on site). I prefer using these to unique visits alone because it hard
      > > > to measure what constitutes a unique visit vs. a non-unique visit.
      > > > Also, the number of unique visits by itself isn't very telling. This
      > > > metric should be combined with average time spent on site to establish
      > > > a better unit of engagement. I will use data from Compete.com
      > > > (www.compete.com) for my analysis.
      > > >
      > > > From the people, visits, and average stay metrics on Compete.com I
      > > > come up with a term called "people hours." This basically serves to
      > > > measure end-user engagement. It is similar to a term borrowed from the
      > > > gaming industry such as the "free-to-play, pay for merchandise" model
      > > > for online games in China. From this engagement metric we can distill
      > > > how valuable a site is based on how valuable the potential audience
      > > > may be.
      > > >
      > > > In this example I use the YouTube's acquisition price and Compete's
      > > > April numbers to set a baseline. Keep in mind that YouTube's traffic
      > > > has grown significantly since its purchase by Google in November of
      > > > 2006, so this may lead to conservative cross-analysis. Nonetheless,
      > > > since Youtube is number one, and number deserves a premium, I assume
      > > > these effects will cancel each other out.
      > > >
      > > > According to Compete.com, YouTube had 150,652,781 visits to its site
      > > > in April. The average stay on its site was 15 minutes and 52 seconds
      > > > (i.e. 15:52) or 952 seconds. We multiply this by the amount of monthly
      > > > visits (150,652,781 x 952) and we get 143,421,447,512 visitor-seconds.
      > > > We can also calculate visitor minutes and hours by dividing this
      > > > number by 60 (which would give us 2,390,357,459 visitor minutes or
      > > > 39,839,291 visitor hours).
      > > >
      > > > Since this is monthly data, we can furthermore divide these numbers by
      > > > 30 to get visitor second, minute, or hour on a per day basis. Doing so
      > > > we get 1,327,976 visitor-hours per day for YouTube.
      > > >
      > > > Now let's divide the price Google paid for YouTube (i.e.
      > > > $1,650,000,000) by visitor-hours per day (i.e. 1,327,976 hours). This
      > > > would give us $1,242.49 per visitor-hour on a daily basis.
      > > >
      > > > Now lets compare some other online video properties. Dailymotion, the
      > > > popular French-based video sites ranks in at number one next to
      > > > YouTube with 23,632 visitor hours/per day or 1.78% of Youtube. Using
      > > > our model this would bring it's valuation in at just shy of $30
      > > > million. Heavy.com, Metacafe, and Break follow respectively.
      > > >
      > > > Here is the the model that I put together in an excel spreadsheet:
      > > > http://edwardjthomas.org/video/video_valuation.xls
      > > >
      > > >
      > > >
      > >
      > >
      > > [Non-text portions of this message have been removed]
      > >
      >
      >
      >

      [Non-text portions of this message have been removed]




      [Non-text portions of this message have been removed]
    • Show all 10 messages in this topic