4 ms·

Why is his web site (Blippy) on a steep decline: http://siteanalytics.compete.com/blippy.com/ http://siteanalytics.compete.com/blippy.com/ ?

pud 16y ago
I'm not full time at Blippy but can say (a) compete.com isn't accurate, and (b) Blippy has various other goings on that aren't reflected there. But that's a different blog post for a different day.
- dennisgorelik 16y ago
  compete.com is not accurate, but it's rarely off by order of magnitude. Are you saying that Blippy is NOT on decline?
  - ivankirigin 16y ago
    Compete is regularly off by orders of magnitude and direction
    - dennisgorelik 16y ago
      Could you give an example of Compete.com being off by order of magnitude? Could anyone here claim that Blippy.com is one of such examples? I'm getting downvoted for expressing my doubts in Blippy, but so far nobody made direct claim that Blippy is not on decline. I would normally use Quantcast, but traffic to Blippy is so small, that it is almost off Quantcast's radar: http://www.quantcast.com/blippy.com http://www.quantcast.com/blippy.com
      - ivankirigin 16y ago
        Tipjoy's embedded js & iframe widgets were measured, not just views to tipjoy.com This made things off by a 100X overestimate. Whoops!
  - code_duck 16y ago
    Compete.com shows troughs, peaks and valleys for sites I've worked for, and the timing appears to have no relation whatsoever to real events and our own data. What it is good for is comparing similar sites.
    - dennisgorelik 16y ago
      Such traffic misreports on Compete are typical for low-trafficked web sites. Can you show an example of a web site that is withing top 20000 web sites on Quantcast.com, but still has an order of magnitude inaccuracies on Compete?
      - code_duck 16y ago
        How would I know the real measured numbers for enough sites to give examples?
        dennisgorelik 16y ago
        If you specify just one example that would be good enough. I can give you my example: http://www.quantcast.com/postjobfree.com/traffic http://www.quantcast.com/postjobfree.com/traffic 355,373 visits per month reported by Quantcast is 10% less than what Google Analytics reports (272,688 Visits per Month). http://siteanalytics.compete.com/postjobfree.com/ http://siteanalytics.compete.com/postjobfree.com/ 126,069 unique visits in March 2011 reported by Compete.com is ~2 times lower than Google Analytics shows for US traffic in March 2011 (252,621 Visits). While Compete.com reports numbers that are two times lower, it shows trends correctly (overall PostJobFree.com growth; drop of PostJobFree.com traffic in December).
gommm 16y ago
Every so often I see people on HN use compete as if it were a reliable measurement... In my experience it's not. Compete shows a few of my websites as declining when they are in fact growing... In one particular case it showed that a website that was tripling in traffic was declining... It's so unprecise that I often wonder if it's not just a random number generator weighted by pageranks and mentions in google trend powering compete, it seems it'd give about the same results... Here's a link from reddit talking about that: http://blog.reddit.com/2010/07/experts-misunderestimate-our-traffic.html http://blog.reddit.com/2010/07/experts-misunderestimate-our-...
- dennisgorelik 16y ago
  Reddit is a good example of pointing out to inaccuracies of Compete.com Still, when reporting on Reddit Compete.com is off by factor of 4 (not 10). And the trend is shown correctly (Reddit growths). For Blippy.com both Compete.com and comcast.com show very small numbers and Compete.com shows steep downward trend. Unless somebody could share more reliable source of data I would assume that Blippy is on decline.

pud 16y ago

I'm not full time at Blippy but can say (a) compete.com isn't accurate, and (b) Blippy has various other goings on that aren't reflected there. But that's a different blog post for a different day.

dennisgorelik 16y ago
compete.com is not accurate, but it's rarely off by order of magnitude. Are you saying that Blippy is NOT on decline?
- ivankirigin 16y ago
  Compete is regularly off by orders of magnitude and direction
  - dennisgorelik 16y ago
    Could you give an example of Compete.com being off by order of magnitude? Could anyone here claim that Blippy.com is one of such examples? I'm getting downvoted for expressing my doubts in Blippy, but so far nobody made direct claim that Blippy is not on decline. I would normally use Quantcast, but traffic to Blippy is so small, that it is almost off Quantcast's radar: http://www.quantcast.com/blippy.com http://www.quantcast.com/blippy.com
    - ivankirigin 16y ago
      Tipjoy's embedded js & iframe widgets were measured, not just views to tipjoy.com This made things off by a 100X overestimate. Whoops!
- code_duck 16y ago
  Compete.com shows troughs, peaks and valleys for sites I've worked for, and the timing appears to have no relation whatsoever to real events and our own data. What it is good for is comparing similar sites.
  - dennisgorelik 16y ago
    Such traffic misreports on Compete are typical for low-trafficked web sites. Can you show an example of a web site that is withing top 20000 web sites on Quantcast.com, but still has an order of magnitude inaccuracies on Compete?
    - code_duck 16y ago
      How would I know the real measured numbers for enough sites to give examples?
      - dennisgorelik 16y ago
        If you specify just one example that would be good enough. I can give you my example: http://www.quantcast.com/postjobfree.com/traffic http://www.quantcast.com/postjobfree.com/traffic 355,373 visits per month reported by Quantcast is 10% less than what Google Analytics reports (272,688 Visits per Month). http://siteanalytics.compete.com/postjobfree.com/ http://siteanalytics.compete.com/postjobfree.com/ 126,069 unique visits in March 2011 reported by Compete.com is ~2 times lower than Google Analytics shows for US traffic in March 2011 (252,621 Visits). While Compete.com reports numbers that are two times lower, it shows trends correctly (overall PostJobFree.com growth; drop of PostJobFree.com traffic in December).

gommm 16y ago

Every so often I see people on HN use compete as if it were a reliable measurement... In my experience it's not. Compete shows a few of my websites as declining when they are in fact growing... In one particular case it showed that a website that was tripling in traffic was declining... It's so unprecise that I often wonder if it's not just a random number generator weighted by pageranks and mentions in google trend powering compete, it seems it'd give about the same results... Here's a link from reddit talking about that: http://blog.reddit.com/2010/07/experts-misunderestimate-our-traffic.html http://blog.reddit.com/2010/07/experts-misunderestimate-our-...

dennisgorelik 16y ago
Reddit is a good example of pointing out to inaccuracies of Compete.com Still, when reporting on Reddit Compete.com is off by factor of 4 (not 10). And the trend is shown correctly (Reddit growths). For Blippy.com both Compete.com and comcast.com show very small numbers and Compete.com shows steep downward trend. Unless somebody could share more reliable source of data I would assume that Blippy is on decline.