jump to last post 1-5 of 5 discussions (14 posts)

Hub Scores & Distribution

  1. Paul Maplesden profile image90
    Paul Maplesdenposted 2 years ago

    Following on from the recent blog post about hub scores and how they are changing, I wanted to do some analysis, mainly to satisfy my own curiosity as to how many hubs are reaching a certain hub score.

    I created a standard deviation bellcurve, based on the bellcurve shown in the blog post; for the stats fans out there, the average score was 61 with a standard deviation of 14.

    You can see the bellcurve below, although it's not identical to the one in the blog post, it's similar enough for some data to be drawn.

    http://s2.hubimg.com/u/12241633_f1024.jpg

    As you can see from this, hub scores range from around 17 up to 100, with the bulk of hubs scoring between 47 and 75. Based on this bellcurve, we can work out how many hubs are scoring at, below or above a certain threshold:

    http://s2.hubimg.com/u/12241637.jpg

    If the figures and bellcurve are too small to read, you can see them in a Google Doc that I have shared here: https://docs.google.com/document/d/1B0- … sp=sharing

    So, what does this tell us? Here are the main takeaways:

    - The average score for a hub is 61
    - There are just over 1,050,000 hubs that have been assigned a hub score
    - Based on monthly page views via Quantcast (28 million), if these were distributed to each hub evenly (which of course they are not) each hub gets on average 27 views a month; obviously, the page views are going to be skewed towards higher scoring hubs, but I don't have that data to hand
    - We can also estimate the percentage of hubs that get a certain score or better; here are the key thresholds

    - 77% of hubs score 50 or better
    - 51% of hubs score 60 or better
    - 25% of hubs score 70 or better
    - 8% of hubs score 80 or better
    - Only 1.5% of hubs score 90 or better
    - Only 0.4% of hubs score 95 or better

    Please feel free to add your own thoughts.

    1. colorfulone profile image89
      colorfuloneposted 2 years ago in reply to this

      That is interesting to see, and helpful to see how our hubs are averaging.
      You must be a brain Paul, that is impressive. Bravo!

    2. YogaKat profile image91
      YogaKatposted 2 years ago in reply to this

      Thank you Paul for being a brainiac and all your hard work!  I am NOW feeling pretty happy with my hub scores.

    3. Lorelei Cohen profile image83
      Lorelei Cohenposted 2 years ago in reply to this

      If that is an accurate reflection of Hubpages ranking stats then either articles here are crap or Hubpages should improve their ranking system to better reflect a fairer bell curve rating system.

      Just my opinionated opinion.

    4. Ben Aidoo profile image75
      Ben Aidooposted 2 years ago in reply to this

      Two things stand out from the comments on the Hub Scores Distribution. Either articles at HP are crap according to Lorelei, or the Hub Scores are nothing more than readers response according to Relache. My view  is that you may have the most stellar hub in town, if it attracts fewer readers you would have a low score. And so what's the relevance of the Hub Scores? To motivate you to attract more readers, that's what it's all about.

      1. colorfulone profile image89
        colorfuloneposted 2 years ago in reply to this

        That is a good point and most likely right, Ben Aidoo.

      2. Lorelei Cohen profile image83
        Lorelei Cohenposted 2 years ago in reply to this

        I really have no idea why traffic was taken out of the equation. It is an important reflection on how well an article ranks on the internet overall. It seems that Hubpages will now be using a system which can be very open to personal bias. Most people largely write to earn just as Hubpages is here to earn or I am sure they would not be here. Targeted traffic generally equates to incoming revenue. Like I said at the start of this comment, I just have no idea why that traffic will now largely be removed from the ranking equation.

  2. Marcy Goodfleisch profile image95
    Marcy Goodfleischposted 2 years ago

    Great information, Paul!  What's the math and Hubscore version of the phrase: "Steely eyed missile man," as used in Apollo 13?

  3. Paul Maplesden profile image90
    Paul Maplesdenposted 2 years ago

    Thanks folks - Although really, I was just being a math nerd to avoid starting on my taxes today...

    1. aesta1 profile image90
      aesta1posted 2 years ago in reply to this

      Paul, this is because, as you've said earlier, the hubs are assessed relative to other hubs but if they are assessed relative to set standards of quality, this distribution will look different. If we are assessing relative to other hubs, this is useful in assessing how my hubs are compared to the others here in HP but is that a good measure of quality given that the real is out there in the much bigger world of search? Can HP use standards of quality and see how our hubs meet these?

      1. Marcy Goodfleisch profile image95
        Marcy Goodfleischposted 2 years ago in reply to this

        Check the post on the admin thread that has the updated graph - it shows the new distribution.

      2. relache profile image87
        relacheposted 2 years ago in reply to this

        HubScores do NOT mean anything outside of HubPages, and are not an accurate indication of writing quality at all.

        They are a measure of reader response to your content more than anything else.

  4. Matt Easterbrook5 profile image82
    Matt Easterbrook5posted 2 years ago

    Good information to know. Thanks for doing the research and sharing that Paul.

    Matt

  5. janshares profile image91
    jansharesposted 2 years ago

    Thanks, Paul, useful and informative.

 
working