Welcome to the Q&A Forum

Thank you for your question! There are many possibilities.

The first and most obvious possibility is just the shape of the model itself. Depending on how the models are stretched onto the 0-100 scale, it is possible that a DA 10 is essentially equivalent to a DR 26.

However, Moz Domain Authority and Ahrefs Domain Rating are two similar metrics but differ in a few critical ways. The most significant difference is this...

Moz Domain Authority predicts the likelihood that your domain will rank in Google based on your link profile, Ahrefs Domain Rating reports the strength of your link profile.  I know this sounds like we are mincing words, but it actually matters quite a bit. Ahrefs might report a higher DR because a site has a lot of links while Moz might report a lower DA because we determined that the specific profile of links held by that site does not correlate with ranking well.

yes

Building links from high DA sites actually creates an unnatural pattern that is easy for our algorithm to detect (and for Google). Now more than ever, you should be focusing on getting links from sites that are likely to send you traffic because they are relevant and high quality.

This is true. The average site will lose 6.1% as the new model "re-centers" Domain Authority. Thus, if you drop just 2 or 3 points, you really have nothing to worry about.

However, there are some industries which will be hit very hard - like link networks and link sellers. These sites can expect to see enormous drops in DA.

This is definitely something we are considering, although we also want to be able to make updates in response to Google updates.

There are, of course, multiple meanings of updates...

1. Updates based on new link data (this happens every other day)
2. Updates based on retraining the model with the same variables (TBD)
3. Updates based on adding new variables (TBD)
4. Updates based on the model itself (TBD)

We are working on determining what cadence would be appropriate for each.

Unfortunately, we will not be including any kind of disavow. First, to be clear, spam score is not based on links but based on on-site factors. The way spam score influences your DA is getting links from lots of bad spam score sites, not your own spam score.

That being said, we can't offer a disavow because we can't confirm that Google has the same disavow. Otherwise, it would make it very easy for individuals to manipulate their Domain Authority.

However, as for negative SEO, our model DOES NOT use penalization. It learns what patterns of links correlate with rankings, nothing more, nothing less. Thus, negative SEO will only impact your DA insofar as it impacts your rankings.

Moz's Spam Score is actually unrelated to links. It is wholly calculated based on on-site factors.

The alt text ought to describe the image. Stuffing keywords into the alt-text is unlikely to have any kind of meaningful impact on your rankings.

The Related Topics model in our various content auditing tools finds and recommends topics based on the content of the page. It uses a single model that, IIRC, is built on finding noun phrases and their close relatives.

In the keyword explorer tool, we offer a combination of several types of related keywords. We offer standard phrase matching, real-time suggestions scraping, similar SERP analysis, and finally a related keywords based on Word2Vec modeling. The "Based on Closely Related Topics" drop down filter in Keyword Explorer is the closest algorithm to the Related Topics you find in our content auditing tools, but they are different in a number of ways, not the least of which is that one is based on a seed keyword and the other is based on the content as a whole.

Is your goal to create a standard HTML sitemap, an XML sitemap, or both?

Hi LAJ,

I am Russ Jones, Principal Search Scientist at Moz, and I am primarily responsible for our keyword volume model. This is a common question we get at Moz and there are typically 3 parts to the issue:

1. Google Keyword Planner combines similar words together for volume. For example, take the term "stepping stone molds". Google actually combines the volume of at least 4 keywords "stepping stone molds", "stepping stone mold", "step stone molds", "step stone mold". So, when you see Google's numbers for the term, you are actually seeing the combined volume. There are certainly even more variants (misspellings for example) that Google combines into that phrase. Google doesn't tell us, which is why we use clickstream data to build our models against Google volume. We call this the "disambiguation" problem.

2. Our model is intentionally conservative. Our models are intended to give users a conservative estimate so they don't over-estimate potential traffic and then waste resources (dollars, time, etc) on terms that just don't deliver the traffic. We would rather you be pleasantly surprised with the results of your campaign than underwhelmed and, even worse, under-water, so to speak. This is accomplished a number of ways, one of which is removing outliers more aggressively if they are greater than the the predicted volume. This also helps with severe disambiguation issues like Google Keyword Planner telling you a keyword like "facbook" is searched tens of millions of times because it considers it the same as "facebook".

3. Our model isn't exactly average search volume. The technical definition of our search volume range would be "the highest range of volume which you could expect 6 or more months out of the year to receive that amount or more impressions". This means that things like seasonality (which definitely impact landscaping related terms) will cause a lower volume average.

All this being said, we are constantly revisiting and improving the model. I am looking forward to an early 2018 revamp of the model which better integrates seasonality. However, in general, our numbers would and should always be lower than Google's because we disambiguate terms rather than join them together.

Not sure this is the forum for that question. Most of us here frown upon the technique because it doesn't actually add value to the web. Is your goal to produce content for your own site? Or is it to get away with spreading duplicate content across the web, like a graffiti artist who always misspells his name hoping no one notices it is always him? Does this sound like a web you want to use?

Anyway, the biggest problems with content spinners is they tend to rely on 2 features, both of which aren't very good...

1. Exact synonym substitution: Basically, the number of variations you get is dependent upon how loose you are willing to be with substitution quality. Either you get a few variations because you are careful (and quality remains decent), or you get a lot of variations and the quality is weak.
2. Sentence structure modification: Switching passive/active, moving prepositional phrases, etc. can provide greater variation but the sentences can often end up sounding quite awkward.

There are some other technologies out there which do a slightly better job. However, all of these end up balancing uniqueness against readability. **Your best solution is always producing unique content by hand, targeted to the audience you are reaching out to. **

Luckily with the Google Hummingbird update, Google is better able to map together phrases with identical search intent. "Car Accident New York" and "Car Accident in New York" return very similar search results, and you will see this across the board for very slightly modified long tail phrases. I would use the words and phrases in a syntactically and grammatically correct fashion, rather than focus on getting every variant.

This IS a question for Search Engine Optimization. Many web services out there will not properly handle UTF-8 characters and muck up the URL, causing it not to work at all and preventing users and Google from reaching your site via the link placed on another site. You might have already experienced this when trying to sign up with your email address while using the international characters codes. It is my recommendation that you link build with the version not including the “ö” and make sure your site has the correct canonical tag pointing to the version with the “ö”.

I think it is important here to note something that is a bit of a misnomer in the industry.

Page Authority (PA) and Domain Authority (DA) are NOT measures of how much a page or domain will benefit you were you to receive a link from it. They are machine-learned keyword-agnostic predictive metrics for rankings. That is to say, a high DA domain or high PA domain are more likely to rank, but they are not necessarily more likely to help your site rank were you to acquire a link from that site.

We know that the most important raw metric still continues to be the number of root linking domains. We also know that nearly all published link graph calculations look at metrics like the number of links per page to divide link value passed from one page to another. Subsequently, the actual value you get from a link can be wildly different than the PA or DA might suggest.

I would take a different approach altogether. If you spend even a second considering whether to get a link or not because of its PA or DA, you have wasted that second. While using PA, DA, MozTrust or MozRank to sort opportunities might be useful if automated and asynchronous, you should never stop and sweat whether it is worth going after a link. Of course it is. Get that link. Go get it now.

Happy link hunting

I can't begin to describe how awesome the Big Data team has been in nailing things out and all the surrounding support provided from across Moz. Things are definitely moving full steam ahead. Great work!

Seeing as you already enjoy 2 top 10 rankings (the profile page and the home page), it is going to be difficult to secure much more as you are fighting against Google's domain clustering algorithm (Google doesn't like to only show results from a single domain, they want some diversity). However, ranking #150 does sound a little crazy.

I noticed a high % of inbound links to that page have been nofollowed. The primary link source was a pr newswire release (https://www.hungarianbiotech.com/gate/v3/news_details.php?news_id=80137&page_id=?next=1641)

Sounds like this page needs some decent links. Are there any organizations, charities, etc. which he is a member of that would offer a link to his profile?

This could be part of the culprit, but generally Google is pretty good at giving domains a "benefit of the doubt" so to speak when it comes to homepage redirects. You should first figure out why the redirect from the non trailing-slash version of your page is a 302 and not a 301. This would be something likely set at the webserver level rather than in Yoast or Wordpress, and it ought to be a 301.

This is the code to make sure a trailing slash is in place for internal directories

RewriteCond %{REQUEST_URI} !(/$|.)
RewriteRule (.*) %{REQUEST_URI}/ [R=301,L]

Did you notice if the same problem is true for directories as well as the homepage?

Unfortunately this is a no brainer. You have to go with Google Analytics. There are so many things that make GA the go-to...

1. Best in class event tracking: Knowing when and what a person clicked on is actually not the easiest thing to do.
2. Incredible data views: You can slice and dice your data in tons of different ways.
3. Tons of community support: No other product has as many free advice and support. It is also easy to find pros to help.
4. Integrations: The Analytics API will integrate into tons of great tools, like Moz, for more insights.
5. Free

The trade off is, of course, that you have to give your user data to Google.

1. The anchor pages aren't going to be indexed separately. If you are lucky, you might get a rich snippet from them in the SERPs, which would be nice. You can see an example of this if you search Google for "broken link building" and look at the top position.

2. Google likely has a crawl budget for sites based on a number of factors - inbound links, content uniqueness, etc. Your best bet is to make sure you have a strong link architecture, a complete and updated sitemap,  and a good link profile.

3. Google can't index the whole web, nor would they want to. They just want to index pages that have a strong likelihood of ranking so they can build the best possible search engine.