Official Update to the Hubscore - what's the likely impact?

1. 69

   makingamarkposted 9 years ago

   Which way do you think your Hubscores will go with the NEW change to the Hubscore algorithm?
   See "An Update to HubScore" http://blog.hubpages.com/2015/02/26/an- … -hubscore/
   
   Note that the revision will:
   * deemphasise traffic - from all sources
   * put more of an emphasis on quality as measured through reader consumption and interaction
   
   I'm guessing that the latter will be measured by amount of time spent on the hub, bounce rate and the clickrate eg number of clicks on links and/or sales made - but I could be wrong. My guess is got to be something measurable that's not a qualitative assessment per se. 
   
   I'm hoping - and would like to see some reassurance - that the revised algorithm is not open to manipulation. We found on out on Squidoo what happens when people start trying to manipulate click rates and the algorithm MUST include the necessary protection to avoid such manipulation.
   
   My personal feeling is that it's a mistake not to differentiate between internal traffic and external traffic. Surely it's still critical to the business model that the hub brings in EXTERNAL traffic? 
   
   So long as it excludes all factors open to manipulation by hubbers I'm sure the revised algorithm will be interesting without generating rancour.

   replyreport
   1. 79

      LeanManposted 9 years agoin reply to this

      The really surprising thing that I saw when I read that post was that what I thought were my worst scoring hubs are still above average for HP!!!
      
      Like anything here I guess it will be one of those wait and see moments to see what is going to happen as the changes roll out.. At the end of the day however it is the traffic from Google and the $$ in my pockets that really motivate me not my hubber score or hub scores...

      replyreport
   2. 72

      relacheposted 9 years agoin reply to this

      It's been long-proven that scores do not effect traffic or earnings, so it doesn't really matter which way scores go.  HubScores are only an internal feedback metric so they really don't have much of anything to do with Hub success.
      
      The people griping about HubScores now will most likely still continue to gripe.

      replyreport
   3. 83

      TIMETRAVELER2posted 9 years agoin reply to this

      Having read all of the posts up to this moment about the big change, I am left with a very bad feeling.  Why is it that our work cannot stand on its own merits?  Why must it be "compared" to everybody else's work so that we can have a bell curve, etc.?To me, there is something inherently wrong with this way of thinking.  If I am working for a company along with others who do jobs similar to mine, should I be rewarded based on a compilation of our efforts?  Or should I be rewarded for the quality of my own work?  I must be missing something here, but I have never felt that my level of success should be tied to what other people  do.  Can someone explain this to me?

      replyreport
      1. 72

         relacheposted 9 years agoin reply to this

         You are NOT working for a company.  That's the mistake many people here are making in how they view the situation. 
         
         HubPages does not think of us, the author/users, as employees.  We are using their service, not working for them.  How you look at someone's experience and how you feel about that is very different when someone works for you versus when they do not. 
         
         This entire site is based on competition because it is essentially about advertising.  If you do not want to play that game, you need to take your writing elsewhere.

         replyreport
         1. 83

            TIMETRAVELER2posted 9 years agoin reply to this

            I have never thought of this site as being merely about advertising.  Yes, I know the income arrives from ads that appear from various sources, but I never connected that to the fact that we are here merely to compete with one another.  I do not mind competition.  In fact, I think it is a good thing.  I do, however, mind having to feel that the quality of my work is not viewed  and/or judged separately from the quality of the work of others.  Having watched the team work with authors here in such a nurturing way over the years, it is hard to imagine that all they care about is advertising.  HP IS a company, and while we are not "employees", we do, in a sense, work for them.  I would like to think that they value us more than what you have just said, and frankly, I think they do.  They may need to have a system for determining who is being successful or not, and I understand that.  In fact, I think the very reason that hub scores do not count for earnings, etc, is the fact that they do look at us as individuals.  So, in essence, while what you say makes sense, I do not really think it is the whole story.  Thank goodness!

            replyreport
   4. 92

      cam8510posted 9 years agoin reply to this

      I'll be interested in seeing if fiction and poetry hubscores respond positively to the changes.

      replyreport
   5. 81

      peachpurpleposted 9 years agoin reply to this

      as long as my hubscore stays at 85-95, i am ok with it. I think i need to update my old hubs first

      replyreport
      1. 83

         Marcy Goodfleischposted 9 years agoin reply to this

         Are you talking about the scores on your hubs, or the number on your profile?

         replyreport
2. 87

   neosurkposted 9 years ago

   I have never really paid attention to the hub scores. The individual rankings mattered on Squidoo because they used to determine how much you were paid at the end of the month.
   
   Since HP's hub scores don't correlate to earnings, it is highly unlikely that there is any need for manipulating and artificially increasing them.

   replyreport
3. 0

   calculus-geometryposted 9 years ago

   Scores on HP do not work like tiers/ranks on Squid.  They have zero influence on earnings. As for people who would spend their time manipulating a metric that has no bearing on earnings, I say let them waste their time.

   replyreport
   1. 94

      SmartAndFunposted 9 years agoin reply to this

      +1.
      
      While I'm glad hubscores are getting a makeover to make them more relevant, the main reason I'll be happy with any improvements is if the scores end up being more straightforward and meaningful, and less confusing to the people who worry about their scores. I feel kinda bad for people who spent lots of time and energy chasing a score that doesn't have much to do with anything, trying to raise their scores or achieve a certain score, and/or asking questions about scores in the forums.
      
      Until the scores have something to do with the views I receive from search engines and the amount of revenue I earn, I am going to continue ignoring my scores. I do welcome improvements, though!

      replyreport
   2. 87

      Marisa Wrightposted 9 years agoin reply to this

      +1

      replyreport
4. 76

   Paul Maplesdenposted 9 years ago

   What is also interesting about this is that the hubscores all fit a bellcurve. This means that hubs are being implicitly ranked against all other hubs and assigned a score (otherwise there would be much more 'random noise' in the distribution.) In other words, hubscores are calculated to fit the curve.
   
   That makes sense of course, and I suspect that ranking a hub against all other hubs (using inputs from QAP, comments, engagement, content etc.) rather than scoring it independently, makes sense. I'll do some more analysis of the bellcurve and post my findings.

   replyreport
5. 54

   janderson99posted 9 years ago

   I suggest that Hub Score should simply be the QAP score, purely and simply! Tell it like it is!

   replyreport
   1. 94

      SmartAndFunposted 9 years agoin reply to this

      Sounds good to me. Why make people feel as though they need to spend time on "analysis of the bell curve?" I guess some people enjoy analyzing stats and don't see it as a waste of time, but it seems like the scores are a little too complicated if people are going to that kind of trouble, especially when the scores have no impact on views or earnings.

      replyreport
   2. 83

      Marcy Goodfleischposted 9 years agoin reply to this

      Janderson - the problems with using the QAP are that not all hubs have been through that process, and no matter how they've tried to engineer the QAP, it's still subjective and inconsistent.
      
      We all see hubs that pass the QAP and have violations, and we all know there's quite a bit of very old and very bad content on the site.  I would be completely in favor of a system that rates things on quality, but I still see things that don't (IMO) meet quality standards.
      
      The point Paul M. made in this thread about the bell curve is very significant. The curve reflects the current scores (which are based on - heck - we don't really know).  It would be interesting to see a similar post with the new distribution of scores across the site, once it is calculated.  I'd be curious if a new chart would show the distribution as compared to some significant (quantifiable) elements of quality. 
      
      I wonder if backlinks are factored (I'd imagine yes - but are they filtered for quality?).  Also wonder what in the HECK they mean by the vague 'consumer' factors.  And if all of that can be artificially manipulated.

      replyreport
      1. 84

         psycheskinnerposted 9 years agoin reply to this

         QAP my be unreliable but the fate of our hubs relies on it, unlike the prominent hub score which is similarly erratic.  So I think knowing what it is makes more sense than just trying to guess via indirect measures.

         replyreport
      2. 54

         janderson99posted 9 years agoin reply to this

         Marcy - so how does all the obscure mumbo jumbo of other factors make it better? "it's still subjective and inconsistent" applies to the current scores. Nothing is perfect, but using QAP makes it clear how it is derived.
         I don't understand Paul M point about the bell curve "hubscores are calculated to fit the curve". Most large data sets fit a bell curve, normal distribution, for scores within bounds (ratings 0-10) - The Good , the Bad and the Ugly.

         replyreport
         1. 83

            Marcy Goodfleischposted 9 years agoin reply to this

            The current curve measures hubs only against what is on the site (not sure what it measures - the invisible cat color Psychskinner meantioned is accurate).  This means there is not an objective score based on known standards of quality.
            
            If the new system is set up to provide true information about quality, then writers can (as suggested) concentrate efforts on hubs that have lower scores. Many writers may never look at their accounts or scores, but this might provide a means for those numbers to have utility and trigger improvements.
            
            The current filters include 'traffic,' so a hub can be of high quality, but have a lower number, which might lead writers to waste time trying to improve 'quality' on a product that's already good.

            replyreport
            1. 54

               janderson99posted 9 years agoin reply to this

               QAP IS supposedly  "an objective score based on known standards of quality". The raters are periodically re-calibrated and standardised and I think several different scoring mechanisms are used. Using QAP would mean that the score remain unchanged, unless the hub is edited or the ratings are changed. It is the wild fluctuations in hub scores over time which disturbs many people and destroys its credibility. Having a score based on QAP would mean that writers could edit hubs and get feedback on how their changes improve the scores. Currently when you edit a hub and add all sorts of stuff the hub scores often go down. You need 'rime and reason' and sensible feedback, to learn what makes a difference. If QAP has no credibility or reliability then what's the point having a score.

               replyreport
               1. 76

                  Paul Maplesdenposted 9 years agoin reply to this

                  I suspect though that the fact that scores are ranked on a bellcurve is part of the reason that hubscores fluctuate. In other words, your hub is being continually compared to all other hubs, based on certain measures, and the score adjusted up or down so it fits the bellcurve.
                  
                  For those that don't quite 'get' the concept of grading on a bellcurve (and it is a strange concept if you've not come across it before), here are a couple of examples (based on grading test scores, but the same principles apply.)
                  
                  http://en.wikipedia.org/wiki/Grading_on_a_curve
                  
                  http://www.wikihow.com/Curve-Grades
                  
                  The key takeaway though is that hubscore isn't based on your hub's absolute, individual merits on its own, but rather on how it compares to all other hubs.

                  replyreport
                  1. 53

                     paradigm searchposted 9 years agoin reply to this

                     Yep. And as junk hubs continue to be eliminated, mid-range hubs will automatically drop in score. Not necessarily a bad thing, a little competition never hurts.

                     replyreport
                  2. 54

                     janderson99posted 9 years agoin reply to this

                     Paul - how do you know the hub scores are graded and actually changed? Traffic has been the major reason for score changes in the past. Curve grading is a 'different kettle of fish' and the marks themselves are not changed - only the grades (A,B,C,D,E and F) are normalised.
                     
                     To quote from one of your examples. "What if, for instance, on a particularly difficult assignment, the few high scores are in the 80% range, the mid-range scores are in the 60% range, and the low scores are in the 40% range? Do the very best students in your class deserve low B's and the average students deserve low D's? Probably not. By using a bell curve grading method, you set the class's mean grade as a middle C, which means that your best students should get A's and your worst students should get F's, regardless of their absolute scores." But the scores or marks don't change, only the grades.
                     
                     There is merit in using absolute scores as you can see the outcome of various measures HP undertakes to improve quality, and how the actions of the authors themselves improves their absolute scores. Having a Quality Hub Score that changes is nonsense, if the hub is not edited.

                     replyreport
                     1. 76

                        Paul Maplesdenposted 9 years agoin reply to this

                        I don't know it for a fact - It's just my intuition, based on the fact that they fit a bellcurve. There may be other algorithms or factors playing a part aside from grading against other hubs.
                        
                        I just find it philosophically interesting that hubs are (very-likely) graded against each other; that's not to say that individual merits don't play a large part (since that will naturally make your 'grade' higher, which would translate to a higher position on the curve), but I do think Paradigm made a good point when he said that as low quality hubs are eliminated, scores of the ones above will likely decrease slightly as things shift around.

                        replyreport
                        1. 54

                           janderson99posted 9 years agoin reply to this

                           I would expect the QAP scores to correspond to a normal distribution, bell curve, without grading when converted to a percentage score.

                           replyreport
                           1. 76

                              Paul Maplesdenposted 9 years agoin reply to this

                              It's possible, although I wouldn't have thought it would be so smooth. Perhaps a staff member could confirm?

                              replyreport
               2. 84

                  rebekahELLEposted 9 years agoin reply to this

                    If scores are a measurement of quality, there is still no feedback or reason why an edit causes the score to drop.  At least if it increases, we can safely assess that the edit was rated higher? 
                  While hubscores are not visible and do not affect earnings, I'm not clear on why they're necessary.  What is their purpose?

                  replyreport
                  1. 54

                     janderson99posted 9 years agoin reply to this

                     Their main purpose is their use to rank pages for listing on the topic pages and on the Related Hubs listings and to contribute to Hubber score, and the many other secret ways HP uses them. The scores are mostly of benefit for HP admin rather than authors, especially when so many obscure metrics contribute.

                     replyreport
                  2. 91

                     aesta1posted 9 years agoin reply to this

                     Rebekah, I also don't understand that and those not edited just somehow retain their better scores so what's the incentive to edit other than your own desire to improve your hub.

                     replyreport
               3. 72

                  brakel2posted 9 years agoin reply to this

                  I improved three hubs recently, spending a lot of time and effort, and the scores went in the 90s.They remained there for one day and two then went backward. Can anyone explain that? I must have a misunderstanding.

                  replyreport
6. 84

   psycheskinnerposted 9 years ago

   I very much like analyzing stats if it is 1) possible and 2) the stats relate to meaningful outcomes.  The Hubscore does neither.

   replyreport
7. 76

   Paul Maplesdenposted 9 years ago

   Don't higher hubscores mean potentially higher placement on topics pages and other places in HP? Does anyone know if there's a direct correlation?

   replyreport
   1. 83

      Marcy Goodfleischposted 9 years agoin reply to this

      Paul - I think they do, but I don't know for sure.  I imagine they'll be a factor in the 'promotion' of Editors' Choice content.  The EC program was supposed to include promotion all along, and it never happened (it was implemented more than two years ago, I believe). So maybe the new scoring system provides an internal trigger that will help in terms of featuring, EC choices, etc. 
      
      Maybe they've developed a way to filter and rate based on word count, number of capsules, photos, videos and the other qualities they've been encouraging. I was so glad when they installed the 'counter' for us to view as we create and edit - prior to that, although I know my content is usually well above 1,000 words, I had to cut & paste into a Word doc to see the length.
      
      It would be wonderful if the new system can somehow discern bad writing. Or copied content. Those are two of the biggest killers on this site.

      replyreport
      1. 84

         psycheskinnerposted 9 years agoin reply to this

         I rather suspect Hubscores have no role in EC.  But then who knows?  Its like trying to determine the coat color of an invisible cat.

         replyreport
   2. 94

      SmartAndFunposted 9 years agoin reply to this

      I don't know, but I would be willing to risk the two or three views a year I might get off of those pages. I don't think those pages have much impact at all on my traffic, the bulk of which comes from search engines.

      replyreport
      1. 83

         Marcy Goodfleischposted 9 years agoin reply to this

         I'm with you on that - my traffic is primarily external, other than perhaps from internal linking here and there. I don't know how valuable internal traffic is in terms of Google smiling on us.

         replyreport
   3. 87

      Marisa Wrightposted 9 years agoin reply to this

      It may do, but that's only an indirect benefit because the vast majority of HubPages' readers never browse the topic pages.

      replyreport
      1. 83

         Marcy Goodfleischposted 9 years agoin reply to this

         The topic pages could be better used to drive traffic (IMO) - the front page of this site does not lend itself to 'shopping' for something to read. A directory of topics with rotating hubs featured would perhaps attract more readers from those who land on the front page. I've asked about this for a long while - not sure if anyone has considered it seriously.  Or, maybe there are good reasons not to do that.
         
         Yes, this site is about advertising - that's the life's blood that keeps it alive. Therefore, we need to be competitive as a site, not with each other. The Hubscore information can help us know as individuals where to put our attention.
         
         As for being rated on our own work (or however that was phrased), that's exactly what this will do.  The distribution across the new curve graph simply shows the number of hubs at a given rating point. I have no problem with that (as long as the ratings can be trusted to reflect consistent standards).
         
         For the Hubscore, perhaps traffic from Google does not need to be measured. Google may not care how much traffic it drives here, but it does care about quality. It's doubtful that we would learn anything useful about a hub's quality if we factor in traffic. This is especially true of internal traffic. The goal here is to help raise all quality to levels that Google will rank better on search results.

         replyreport
8. 94

   SmartAndFunposted 9 years ago

   I wish they would just dump the scores.

   replyreport
   1. 91

      aesta1posted 9 years agoin reply to this

      I agree.

      replyreport
   2. 72

      brakel2posted 9 years agoin reply to this

      The site doesn't have many scores overall in the 90s, according to the bell graph. That surprises me.

      replyreport
9. 60

   ciutysmantposted 9 years ago

   I don't get one thing - why my new hubs have higher than average Hub Score? Definitely, I'm happy about it, although it does not make sense...

   replyreport
   1. 72

      brakel2posted 9 years agoin reply to this

      If QP rates them that way, that is the way it is.  70s are in the higher than average.

      replyreport
10. 89

    Millionaire Tipsposted 9 years ago

    A while back, I unpublished a whole bunch of hubs. These were hubs with low scores, and no traffic. They didn't really seem worth the effort to try to fix.  Now, they have higher scores than my published hubs!

    replyreport
    1. 83

       Marcy Goodfleischposted 9 years agoin reply to this

       Mine are going crazy, too, MT - it feels like I'm in an ER with an unstable patient whose blood pressure keeps fluctuating dramatically. No changes in traffic, and it doesn't matter whether they're published or not.

       replyreport
       1. 91

          LongTimeMotherposted 9 years agoin reply to this

          I also have erratic changes in score on individual hubs, despite their traffic levels remaining relatively constant. Some hubs have a drop in traffic so it is probably time to check for stolen copies and file more DMCAs.

          replyreport
       2. 89

          Millionaire Tipsposted 9 years agoin reply to this

          It makes me wonder if those were worth salvaging, but I am going to wait until things stabilize.

          replyreport
11. 91

    Kylyssaposted 9 years ago

    I don't think HP realizes that people will try to use something that looks like a grade or a percentage as if it were feedback. Nor do they appear to realize that undependable and erratic feedback is actually worse than useless; it's misleading and discouraging.
    
    Imagine a high school student getting homework and papers graded like hubs appear to be graded. The student works on polishing a paper, hands it in, and gets a grade of 88. Students are then given the ability to edit and improve the papers they've just turned in so the student, taking pride in his work and eager to do better, gets right to work. He researches his topic, fixes his grammar, tightens up his text, adds fifty percent more material to his paper, and gets a straight-A student to look it over. He follows her advice and then turns in his paper. Within hours, he gets his paper back, but now his grade is only 70. The teacher tells him to learn basic English, fix up his paper, put some effort into it, and he'll be able to get a passing grade on it. So the student researches more, writes more, and hires a college graduate student with an English major to edit his paper. He turns it in again and gets it back with a score of 65. He then learns his papers are being graded by people for whom English is a second, third, or fourth language who have been hired for the job without passing an English proficiency test.
    
    I don't know about you, but that would drive me batspit bonkers.

    replyreport
    1. 60

       jodijoyousposted 9 years agoin reply to this

       +100

       replyreport