Update: Why did my Domain Authority change?
In this index update we re-calibrated our Domain Authority metric to better reflect the relationships between all domains on the Internet. This means that many websites' Domain Authority (DA) changed.
Not to worry! If your domain authority went down, so did all of the other domains that had similar link profiles before the index update. (Don't think about it like something bad happened to your site, think about it like we changed how we view the entire Internet) You can read more about why we did this in the section below.
The good news is, we have an index update for you! And it's a couple of days sooner than previously announced. The bad news is, things were, as you might have noticed, a little rocky this morning. We had more traffic to the API than ever before, and, through the magic of being a scrappy startup, we all jumped into action. Fortunately, through the magic of Amazon Web Services, we've quickly increased our infrastructure and are serving better than ever. I do apologize for any issues this might have caused.
I've got a few things to say in this post, so you can skip forward if you like:
- Page Authority and Domain Authority Change
- New API Dashboard
- Stats on Nofollow vs Rel=Canonical
- Index Update Stats
Page Authority and Domain Authority Change
We've gotten a lot of feedback about Page Authority and Domain Authority. We're excited about these metrics and are using them to power a lot of what we do: sorting links, crawl selection, keyword difficulty. But lately, it seemed as if things were getting a little... clumpy. We were packing our numbers too closely together to give a real sense of the spread in authority over the web. I'll defer to Ben and Rand, who are working on this a lot, but just to give you a taste:
.png)
As you can see, we've pulled apart a lot of great sites. This spread, for example SEOmoz with a PA of 80 and Amazon.com with a PA of 89, better reflects different authorities.
This will have effects across tools, including Keyword Difficulty. So take a minute to check those out and make sure that what we're showing you matches your intuition.
It's actually very fitting that we should have more traffic than ever before, because we've been hard at work on better serving one of our biggest API consumers: You! Today we're launching our SEOmoz API Dashboard.
.png)
This dashboard will be the place to go to manage your SEOmoz API account. Right now we're including all of your API usage. This gives you visibility into your API consumption, critical if you're on the paid plan. And if you're on the free plan this gives you some idea of the usage of your tools. As we improve what we offer both in the API and to support application development, you'll see more and more here.
Stats on Nofollow vs Rel=Canonical
Last week I had a great chat with Eric Enge at Stone Temple Consulting. We talked a little bit about the usage of nofollow and rel=canonical over the last year (a big year for both!), but I didn't have anything concrete to share at the time. I dug into it and it's pretty interesting:
As you can see, rel=canonical is really taking off. Since we started keeping really good stats on its usage in our data, it's grown in usage by about 50%, in just six months! We see rel=canonical being used more than either internal or external nofollows. And internal nofollows have fallen off quite a bit about eight months ago, but are reasonably stable since then.
My hypothesis (without supporting data at the moment), is that two mindsets are winning:
- use rel=canonical right away
- if rel=nofollow is working leave it
I'll leave it to the expert SEOs to debate this (in the comments, please!), but that could well be sound advice.
Here are some charts and graphs of the data we've updated since last month.
We're staying on course with our current update rate for pages. We've got updated information for about 43 billion pages.
And we have a corresponding update for links to and from those pages.
.png)
We've got two focuses for our data updates:
- Get those domains which we used to think about as niche
- Get deep on those domains that are highly authoritative
This is actually a big initiative we've been working on this year. And we're already seeing great improvements to our data quality.
I hope you enjoy the data. As always, feedback is much appreciated!
Great job continuing to dial in the numbers. Of course, that will be constant need as the tool and the web continue to evolve.
Got me thinking, which is never a good thing...
would love to see this evolve to identifying the market segment (industry, topic, general link profile, etc.).
Perhaps even presenting a view like "and these appear to be related, sites within the segment," but more valuable, a relational metric, kind of a barometer if you will... so not just a 80/100 score, but how that relates to the segment... above, below, average, etc.
These are really good ideas. It's nice to hear support for data segmentation. We've gotten that suggestion before, and the more we hear people clamoring for it, the more likely we are to invest in it.
Hi,
Thank you for the hard work and free tools.
Some comments regarding the new PA (Page Authority) and DA (Domain Authority) numbers:
The new numbers have much higher discrepancies where the home page of a web site can have a much much higher PA than the DA. Take a site like stanford.edu for instance where the home page PA is 91 and the site DA is 54.
These new numbers make the tool a lot less useful as it's much harder to gauge how strong a site really is. With the previous numbers a site like stanford.edu had a 100 DA which seems fitting for one of the strongest sites on the net.
I found the PA and DA numbers very useful until this update and together with PR, number of links and some other metrics really gave me the tools to judge a site or page's strength.
Hey Adam - I totally hear what you're saying and the real answer is that we are still working towards being great on this front, but the latest numbers should be a big step in the right direction.
Stanford's homepage should have a really high PA, because it can rank really, really well! However, not all content on the Stanford domain is quite this amazing. There are lots of parts of the site that don't perform particularly well in Google's results, so giving it a massive DA would be irresponsible and inaccurate of us.
PA - predicts how well that individual page can perform in Google
DA - predicts how well any page on that root domain could perform in Google
I hope that helps - certainly happy to continue the conversation.
Hey Rand,
What you are saying makes sense, but my biggest use for DA was for "guessing" overall domain authority. How valuable would a link from that domain be in a seach engine's "eyes" just from an authority standpoint. I always felt that getting a link from a high authority site, even if it's from a low authority page, worked wonders for a website's authority.
The new tool metrics are great for judging page strength, how valuable a link would be from a specific page, but not that great for how valuable a guest blog post would be on that domain for example.
Other than the above I also noticed some results which didn't make sense to me where sites which I know are a lot weaker on the authority front have much higher DA scores than sites with higher authority.
PageRank is not very exact, but it does usually go hand in hand with authority on a very rough scale, usually I've found that at the worst case to a factor of ten. The previous DA numbers were much more in tune with this.
I don't know how much machine calibration was was put into the numbers, but I'm assuming much more from a PA standpoint.
I have to agree with Adam here. I found the previous DA value *very* useful to gauge how valuable a link on a domain would be (much better than Page Rank or any other metric available), and used it daily to help find links worth pursuing, blogs worth guest posting on, etc. Perhaps you can make a new metric - "Domain Value" - that estimates the overall value of a link on the domain? That way, you can keep DA as you intend it, and have a new metric that satisfies these other needs?
Fixed a few typos
Adam & Jeremy - actually, the new DA numbers should follow exactly that same intuition and should be (generally) more accurate than before. For example, SEOmoz, SEObook, SearchEngineland and SearchEngineWatch all used to be clustered between 89-93. They're not much more logically spread out and a distance from sites like Wikipedia (DA 100) that should make more sense, too.
I know it's tough to have to re-calibrate your anchoring on a metric you've gotten used to, but we also felt it was wrong for us to show potentially misleading numbers. I can also say that, as we announced when these metrics launched, we're going to keep tweaking and working to make them better, so more shifts like this may happen (though probably less dramatic in the future).
Watch for a blog post next week that helps show correlation analysis on all the metrics from Linkscape and tons of other third-party sources!
Thanks for the update, but this has coursed a problem!
now my 1 metre rules has been re-callibrated my preference measures are useless. Are you saying that the DA and PA are constantly being re-calibrated or can I now use these measures with confidence.
Chris
SEO TOP PAGE
-Casey Removed Link
we just launched PA/DA a few months ago and this was certainly a big shift. We want to dial that as close to stable as quickly as possible. So there will be some shifts going forward.
But I feel your pain. I'm a big consumer of these metrics in my work too, and this certainly caused some frustration for me. I'm sure the team will work to keep the fluctuations more subtle in the future, but always helpful.
Ah, the value of Amazon Web Services and the materialization of cloud computing, everything you need without:
1) upfront investment in servers
2) room to put servers
3) electricity to cool the room
4) IT people to maintain the servers
Best of all? if you need to scale Amazon can handle just about anything you can throw at it.
kudos for choosing Amazon - smart choice
Yes, we love AWS :)
Really a great job.
Just help me solving a doubt: does the DA / PA calibration meant also a mozRank/mozTrust calibration too? (myown answer is that it seems not).
Yeah - that's right. MozRank, mozTrust, DmR & DmT haven't changed much at all (and the calculations behind them are very similar). mozRank is still well correlated with Google's PageRank.
This is really about PA & DA getting more intuitive and more predictive as we get better and better about determining the importance of sites and pages on the web (and their ability to perform well in Google).
Hi Nick,
Thank you for all the great info about the update (and Rand for covering it last night in the webinar too).
The API interface seems to be down. Can you give a quick heads up about what we can do with the API?
Thanks!
Thomas
can you PM me or send an email to api [at] seomoz [dot] org with more details (including the request you're making)?
It's a big priority for us to keep this reliable.
Update: I have identified the update poitn in my Analysis.
Hi Guys,
Great Work as usual guys however can I ask quite a pedantic question. When exactly (time wise) did this change get rolled out?
I was doing some intense competitor analysis yesterday day/night for a client and think my results may now be a bit screwy.
My Client had a Domain Auth of 85 (captured day before yesterday) which i was using as a benchmark. This has now gone down to 61 which isnt a problem as you say. However when comparing old Domain Auth to Domain Auth of other sites post update this obviously produces funky data.
So rather than redoing (45 pages) of the analysis I was hoping to be able to identify when the update actually took place and just amend accordingly.
Looking at my report now I think I can see where the shift occured but would be helpful if you can be more specific for me.
Cheers!
Sorry more from me! I dont quite understand why the Keyword Difficulty figures have changed so dramatically.
For instance keyphrase "Photovoltaic Panels"
Before Update - Keyword Difficulty 73%
After Update - Keyword Difficulty 42%
I understand that the scale has changed but the competitor websites in question have not nor has their relative Authority.
Is the drop of over 30% indicative of the widening of the scale? Seems like a major leap considering the key term is as competitive today as it was two days ago.
Unfortunately I think you'll have to try out a few KW difficulty runs to re-calibrate your own understanding.
KW difficulty is considering difficulty across the whole web and all queries. So in this case I suspect it's not that difficulty has dropped in this SERP, but rather that we're recalibrating things across the web.
its so awesome that you can show a rise in use of rel=canonical
how many rows does the api return a month?
Hi,
I found a site with PA: 52 and DA: 84. It only has 13 backlinks from 9 domains, all of them with low PA/DA. is this a bug?
can you PM me the page/site?
The link is in my previous post (through OSE), here it is again. Feel free to remove it if it should not be made public.
Funny thing, the homepage has a mozrank of 0.05 .
I knew something was going on before you posted this information, when I saw all of my websites with different PA/DA results yesterday.
Glad to hear about this; Getting more accurate metrics can only helps us to win the struggle with our competitors! ;)
Can't wait to lay my hands on SEOmoz API, as I keep hearing about it a lot in the community... I'm planning some tool development soon... Aaaah If I got more time!
I got one question tough: You want us to debate on rel=canonical vs rel=nofollow, but I don't understand the context... I know rel=canonical act this way , but how can "leave" rel=nofollow on my internal linking? Are you saying that rel=canonical can be a replacement of rel=nofollow? Maybe it's because I haven't had my coffee yet, but I don't understand...
Thanks guys!
You're right, rel=canonical and rel=nofollow aren't competiting, solving the same problem, etc. I'm sorry I gave that impression.
More I'm saying, issues for both of these came out over the last year or so. So what do you think about giving people the advice to invest in altering strategy around one versus spending that energy investing in the other?
Very good job, thanks !
It would seem that the PA/DA results are more approximate for the french results (google.fr) than before your new calibration.
The correspondence PA/DA - position in SERP was much better respected before in google.fr.
You guys are doing some awesome & valuable work here, so thank you.
(Edit: I'm just frustrated actually)
For the first time I have to disagree with that.
I really think that this was the worst move ever.
For the past weeks I've watched the updates and sticked to the numbers closely. All of this now renders my past efforts useless. I tried to keep track of my Index's PA as a good metric for my SEO effort, but all of my previous data is now outdated.
This won't be easy to adapt to the new numbers and way of thinking.
Now, when I look at my website's backlinks in OSE, I see the most powerful link to my website is a nofollow comment from Hongkiat (PA of the page : 67)...
Come on, as SEOs we know that's not as good as the backlinks I had on some of my competitors' index.
But still, thank you for your work and I appreciate a lot (really) the updates :)))
I don't think you should base your opinion on the updates on the fact that it means your data is no longer relevant. They have to update the system. This may mean that there are certain metrics discounted, sites are re-evaluated and that sort of thing.
So rather than thinking 'oh balls, my data is screwed now'; you should be thinking 'oh awesome, I have more up-to-date data that I can work with and improve'.
Thanks for making the change. It makes everything more understandable. I could never figure out how I was a 70 and google a 90.....lol.
Thanks for the hard work on the link update!
I have a question: Once you have indexed a page do you re-index it for new links?
Yeah - we'll revisit links and pages we think are important each month (sometimes more frequently). That's also how we can see that, as Nick has previously noted, 80% of the URLs on the web die/change/redirect within 12 months, and more than 50% within 6 months. The web changes really fast, so we have to keep on top of it (improved freshness is certainly something we're working towards).
So cool that you answered my question so quickly - Thanks!
2 Follow up questions:
1. When building links by commenting on blog posts that are old but have high pagerank do you know guidelines for when google stops recrawling them? 2 years, 3 years, etc?
2. When building links by commenting on blog posts do you feel it is better to have your comment/link in place asap - i.e before google crawls the page so your link is a part of the initial/original index/crawl versus commenting on old established blog posts?
#1 - I doubt they'd ever stop completely. It may just be longer and longer between re-crawls (say 40 days instead of 4 days).
#2 - No, certainly not, but I'm not sure about the strategy overall. Most blog comments are nofollow and many that are dofollow get a lot of spam. I certainly like blog comments as a way to engage effectively in a community and get bloggers seeing/noticing you (and they might link in the future), but just drive-by comments with links may yield very poor results.
You are the man - I owe you a candy bar or something - Thanks!
Just a thought...and it seems to work okay for me. As Rand said, only post on blogs that you intend to become part of the community. When I find a great blog post that is older, I'll still comment on it. Then you can add a social bookmark w/ any of those sites like Digg, Reddit, Propeller, Jumptags, etc. that will sometimes get it crawled sooner than later. Just a tip....seems to work okay for me if the post is older.
Very good idea! Thanks for sharing it :)
Really? Am I missing something? Google.com has a DA of 94, with 724,796 linking root domains and 70,730,726 total links? Is it now impossible to have a score of 100? Facebook is 97... I don't know ~ I understand the scale is logarithmic, but it just doesn't seem right. My Grandma's tea tree oil farm's website, with only 9 root domains and 15 links has a DA of 38 (and the "strongest" link to her site only has a DA of 54). This seems high, and surely can't be right...
The DA scores are predictive of how well a domain overall performs in Google's rankings. Perhaps surprisingly, a lot of pages on Google.com don't do that well in Google's own rankings (Google Finance, Google Profiles, Knol, etc. are all examples). Hence, Wikipedia and Amazon are 100, Yahoo! is 97, Google is 94, Flickr is 84, eHow is 63, etc.
We hope this feels much more natural to how pages on those domains perform in Google's rankings. We'd love to hear/see outliers you find as it can help us make this data even better.
Thanks Rand, that really clears things up for me. Domain Authority is not a direct measure of the "authority" of a domain - but rather an indicator of how well it will perform in Google's rankings. That's what I needed to wrap my head around.
Thanks!
Exactly - if we take every other metric we have around a page/domain and mash them together in a machine learning model that creates hundreds of derivatives and generates complex equations to help predict how a page would rank (or how any page on a given domain would rank) in Google, that's what the PA/DA numbers are.
We really think about them as a more sophisticated way to value pages and sites and to see the level of competition we're up against in the SERPs.
Thanks guys, keep up the great work. We love the new tool and all it's functions.
Hi,
I found a site (www.memocall.fr, www. are important) with PA : 37 / DA : 43 and only 1 link from 1 domain !
(you can remove the URL if necessary)
Adifco - France.
Thanks for the report. We're building a list of these for further investigation. Feel free to PM me with more examples as you find them! :)
awesome ,tks for u
Great news, I'm hoping that someday within the next year or so I'll be able to get my hands on a shiny PRO membership so I'll be able to reap the benefits as much as possible.
Well, congratulations... your avatar says you did it :)
hi,very valuble information.i am new to this search engine optimization.i have question ..when ever i add new page to my site the ranking or index of my site go down.why it happens?
You should check out our beginners guides. Those are great places to start.
And (ahem) I'm sure our community is always glad to have someone new join into the discussion. Usually this kind of question is more appropriate for Q+A, or more relevant blog posts.