Can you trust Domain Authority and Page Authority metrics?

Page Authority (PA) and Domain Authority (DA) are two new link related metrics introduced by SEOmoz early this year (in relation to the Open Site Explorer launch).

PA and DA are said to be query-independent link metrics designed to help compare how “strong” or “authoritative” pages/domains are in terms of ranking potential.

We think Linkscape and Open Site Explorer are both great tools and use them in our everyday work but we are careful not to overly rely on other people’s work without taking the time to ensure it is robust and reliable.

We have there taken a closer look at the Page Authority and Domain Authority metrics and accept the fact that the correlation between Google rankings and PA/DA is not perfect. Rand Fishkin (of SEOmoz) and Will Critchlow (of Distilled) talk about a 70% correlation to Google’s rankings (see video part 2).

Our main concern comes from our findings of inconsistencies and our inability to identify whether these inconsistencies are because we’re in some edge case for one reason or another.

Even if SEOmoz are unable to post mechanics behind the PA/DA metrics it’d be useful to know the situations in which they are most likely to be unreliable.

Not knowing how exactly DA is construed would not be that bad if the results didn’t deviate too much out of what we expect them to be. The problem is, we have seen some pretty weird DA scorings throughout the last months and our trust in the new metrics has been somewhat undermined.

Take a look at two examples – one (what we believe to be) an underestimate and one a severe overestimate.

dacomparison2

Comparison of link-metrics for MyGoudDiet.com and BrisbaneTimes.com.au (Open Site Explorer report)

Overestimate example: MyGoutDiet.com

The link portfolio of this domain can be described in one word: BAD. The website has back-links from 8 domains. All of the links we checked were no-followed (usually blog comment links).
The Page Authority of the main page is 37, while the Domain Authority is a whopping 67! For comparison, SEOmoz.org has a DA of 64, while Google.co.uk has a DA of 50. Sweet, isn’t it?

Here is a list of some of the website’s available link metrics:

  • PR (main page): 0
  • Link count (OSE): 5
  • Linking domains (OSE): 6
  • Linking domains DA (OSE): 52, 52, 49, 48, 41, 39
  • Link count (Yahoo!): 43
  • Linking domains (Yahoo!): 8
  • Percentage of no-follow links: 100%
  • We are not aware of any 301 redirects to the domain

We submitted this example about two weeks ago to the SEOmoz team (we didn’t get back any comments on it though).

Underestimate example: BrisbaneTimes.com.au

The Brisbane Times on the other hand has a quite impressive link portfolio:

  • PR (main page): 7
  • Link count (OSE): 231,114
  • Linking domains (OSE): 798
  • Link count (Yahoo!): 2,280,000

What is its Domain Authority score? 39.

This seems way too low, but can partly be explained by the fact that PA/DA is supposingly not that accurate for local (non-US) websites. When compared with a larger array of Australian websites, the DA score still seems strange.

In conclusion

We’ve made a few more guesses as to why this might be happening, but they are just that – our guesses. Until we get some meaningful feedback from the SEOmoz team we’re a bit sceptical of using the new OSE metrics in everyday work.

Have you been using the Domain and Page Authority metrics? If so, we’d love to see your opinion in the comments below.

TwitterLinkedinGoogleplus

Nemek takes care of SEO at Matter Solutions. He occasionally tweets here. More posts »
 

5 Responses to “Can you trust Domain Authority and Page Authority metrics?”

  1. Suisse Gold

    One site I noticed on OSE…wwwr.creditsuissegold.org

    Linking Root Domain: 1 – Itself!
    Total Links : 12 !!

    Domain Authority: 100!!!!

    Reply
  2. jose leite

    I do not use those metrics to check the links. I would prefer to check indexed pages and PR instead of evaluating the DA or PA. Also check the inbound links in Google (also not perfect) but at least you’ll not live with a maybe wrong data info.
    Jose Leite

    Reply
  3. Nemek

    I agree with you Jose that looking at raw link-related metrics seems to be more helpful in many cases. I’ve found the number of linking root domains to be a useful indicator (in combination with Page Rank data it can be used to spot sites penalized for link selling).

    Reply
  4. Royur

    Hmm… When I check the data, its different from yours…

    My Linkscape data show mygoutdiet.com DA: 11 and brisbane.com.au DA: 79

    Reply
    • Nemek

      Hi Royur, thanks for your comment.

      You’re right, we received a message from SEOmoz that they have looked into this and apparently the scores have been updated now.

      Reply

Leave a Reply