Any SEO moz users notice a HUGE change in OSE (Open Site Explorer) link data numbers?
-
Hi All,
I am having some serious concern with OSE data recently for numerous clients, one client I want to talk about today has the following data from OSE for the month of August 2011 compared with July 2011:
Total links to the domain: (decrease of around 100,000+)
External Followed links: (decrease by around 5,000)
**Linking Root domains: (decrease of over 60) **The crazy thing is that the domain authority has actually gone up by around 5 points for this client even though every thing has suddenly gone down?
Also funny thing is we have been link building quite strong for this client over the last 12 months using only high quality sources from out niche.
I am worried that their is serious issues with the data, I realise we saw some updates to OSE recently yet I am suprised it can be this drastic.
Kind Regards.
PSV
-
This topic affects everyone and Rand offered a specific reply to address these concerns. In short, they are working on the issue but there is only an opportunity to make adjustments once per month. Then the crawler is set to do it's job and the results wont be seen for another month.
The crawler has to go out and pull data from literally billions of web pages. The data then needs to be processed which takes 1-2 weeks. The team is also trying to improve the quality of the tool as well. My recommendation is to use OSE as a guide, not a solid list of all links.
-
Almost 2 months later from last post (Sep 5) here and I am wondering why this problem with OSE is not rectified??? Oh yea, here's another $100. Thanks so much!!?
-
Hey PSV,
Sorry I'm jumping in so late to the game on this! The response from the Customer Service team by Kenny below is correct - with the adjustments we've made to the new crawler, we're going less deep this crawl than we did in the July crawl. This is because we're still trying to make tweaks to get the domain diversity back up to what is what before. It looks like you had a major change from July to August which is strange - I haven't seen such a drastic change pointed out by any other users.
If I could get the actual domain, I can look into this for you. Is this discrepancy being reported by Open Site Explorer or are you querying the API directly?
It's hard to speculate without knowing which domain you're referring to, but you may see a decrease in inbound links if you had a lot of links from a high MozRank page. If you had a large amount of links on one domain and we didn't crawl that domain as deeply this past index where we went MUCH deeper the index before, many of the links deeper down in the domain won't be in this latest index.
Please let me know which domain you're referring to and I can try to track down some information for you!
Thanks,
Cairn
-
I have seen the same thing and posted a private Q along the same lines, having seen substantial changes with a couple of site profiles, both of which have had link building done very carefully with high value sources. Majestic and Google WMT still show the links - just a spanner in the works at the mozplex by the look of things until the next crawl.
Response as follows:
| Kenny Martin | Aug 31, 2011 |
| Customer Service Expert at SEOmoz |Sorry that you're seeing such drastic changes with the new Linkscape update. We've been adjusting the way that we do our crawls over the past couple of months, and the updates made in July were a bit off the mark. Our new crawler was built to be fresher, but it was going deeper into domains, and, unfortunately, not visiting as many domains. Domains with a high MozRank were getting crawled deeper, but domains with middle to lower MozRanks were not getting crawled. With our newer Index, we tried to get the domain diversity back up to or better than that of our last update (as was originally designed) so we didn't crawl as deeply into some of these higher ranking domains as we had for our July index.
Also, keep in mind that it can take some time for links to be included in our Linkscape index because it takes a few weeks to crawl the web and a couple weeks to process all the data that is acquired.
I hope that helps clear things up and we are anticipating on improving our crawler and index each month. My apologies for any inconveniences
-
I'm sorry PSV, I missed the 60 root domain information which you clearly included.
100,000/60 = 1667 so it sounds like some if not most of the links lost may have been site wide links. It's impossible to tell with this information alone. There are too many factors involved:
-
how many domains are linking to your site
-
which domains did you lose? Were they high DA sites or low?
-
which links did you lose? Were they foot links? Links in content? Links from high/low PA pages?
-
were they links to your home page or deep links?
-
did the links use anchor text or simply provide a URL?
-
How is your site's traffic? Have you noticed any recent changes?
-
How many total domains link to the site?
Do you save your links in Excel or have a tool to compare your current links with the past links so you can determine which ones have been lost? This step is key to understanding how to approach the change.
-
-
Hence why I noted 60 root domains loss, which I find a bit weird, as our other software shows gains.
I thought it must have been down to a system update!!
But yeah easy as to get a footer link a partner site footer, but it is not quality
-
Hi Nigel,
Yes we are using several tools for comparison currently we do not limit our selves to only SEOmoz data. I read Rands post about the link scape update, saying how lower quality smaller sites were removed from the OSE index yet this site we deal with has more high quality links in its profile.
-
I agree with Nigel's assessment and would add one thought. Look a lot closer at the number of linking domains, not the total number of links.
You can add a single footer link to a forum site with 100k pages. The link's value is minimal but you would notice 100k more links. What can you do with that information? Very little. It distorts your true backlink profile. If you were to lose that footer link but add 1 quality link from a source such a the New York Times, your sites DA may increase even though you are technically down 99,999 links.
-
I saw similar changes to the Data sets in the August update. I received feedback from moz, was that with the July data set they were crawling deeper into some of the larger (and more trusted) domains which was, in some cases, producing unusual results. For example, we had links coming from domains such as Warnerbros.com, which have no trace of a link to us.
I think the latest update has moved to rectify these issues but at the same time crawled a wider range of results giving a more accurate picture of link graphs for pages. I would suggest that comparing Data month and month to yourself is not a useful as mapping the changes your competitors are having i.e. with Seomoz making the great changes they do to make their metrics reflective of what matters in the SERPs the data will change month on month but if you are looking at how your competitors shift as well you'll get an idea as to how you perform.
I would also suggest using several tools, (whilst linkscape is pretty darn awesome) it's always worth getting a full picture of the link metrics by checking linking domains in webmaster tools, yahoo site explorer, bing tools, majestic seo etc.
In short I don't think there are serious issues with the data but it would be worth checking which domains you've lost and which you've gained, just to be safe.
Got a burning SEO question?
Subscribe to Moz Pro to gain full access to Q&A, answer questions, and ask your own.
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
Are there any strategies (Moz-specific or other) on how to improve the Total Links metric against competitors?
In competitive link metrics, our client is doing very well in every metric against every competitor except in Total Links against one competitor in particular. Wondering how to improve this metric against this one competitor. Any strategies, whether MOZ-specific or other that can boost this metric? Thanks.
Moz Pro | | buffalonickel0 -
SEO moz Report Card
I just ran some on page report cards. As I was playing around with the tool I noticed that I would get different results if I used my primary domain vs a 2nd domain. The main difference was in how the tool was counting keywords on the page. The keyword used was 'vehicle inventory' Primary domain: www.brand-state.com/inventory.htm Title = 1, URL = 0, Meta = 1, H1 = 1, H2-4 = 1 Body =1, Strong = 1, IMG Alt = 1 Total = 7 2nd domain: www.company-name-brand.com/inventory.htm Title = 1, URL = 0, Meta = 1, H1 = 1, H2-4 = 2 Body =5, Strong = 4, IMG Alt = 2 Total = 13 I can understand if the keyword was in the domain, but it's not. So I'm wondering what is going on here - any help or suggestions on what to research would be a great help. Thank you!
Moz Pro | | gormaniavt0 -
"Loading" error in Open Site Explorer
The "top pages" keeps hanging while trying to retrieve social metrics "FB Shares" "Tweets" and "Google +1". I've tried in Chrome and Firefox. Can I just turn these off? ose-hanging.jpg
Moz Pro | | cpaddock0 -
Why doesn't Open Site Explorer recognize the link from my twitter profile?
When I run a report on my competition I'm seeing their link show up in OSE's report under the "Linking Domains" tab, but when I put my site in there, it doesn't recognize my link from my twitter profile. Any ideas?
Moz Pro | | shawn811 -
Links not appearing on Open Site Explorer
My site gained several new inbound links during December and only two of them are not all showing up on the latest Linkscape update. It seems to be the links that were created at the end of the month which are showing up, whereas a handful at the beginning of the month are nowhere to be seen. All the linking pages have been indexed by Google the links are do-follow, and one of the sites in particular is not obsure and has a DA in the 90's. I appreciate the Linkscape doesn't index everything, but I would have thought that more tof the results of my efforts would have shown up in OSE. I'd be really grateful if anyone could explain this to me please. Thanks Ben
Moz Pro | | atticus70 -
Why is Followed Linking Root Domains higher than External Followed Links?
Surely there must be at least one external link for each linking root domain? Some results for smaller sites give a higher number of domains linking in than incoming links - e.g. www.forbesandsawyer.co.uk Under Subdomain metrics: External Followed Links - 1 Followed Linking Root Domains - 2 Surely 2 root domains would mean AT LEAST 2 external followed links? Thanks, Andrew
Moz Pro | | Silktide0 -
Open Site explorer (last 1-3 updates) shows new and wierd results
Hi Moz'ers! I have a questions for you guys. Can you explain why the website "Netspiren" has a lower "domain Authority" than "Duft og Natur"? See screenshot. I know it's a tough question without a detailed analysis, but looking at OSE's data, does the mozrank,moztrust, C-block don't reflect the power of the Domain Authority. What's you take? BTW: "Netspiren" has always had more and better links, and also had a higher Domain Authority. Something about the Domain Authority changed since the past 1-3 OXE updates. Looking forward for your take on this! Have a good weekend! mnRPI.png
Moz Pro | | FrederikTrovatten220 -
Link Count Per Page Including JavaScript Links - Should We Worry About Them?
With large ecommerce sites, we usually have more than 100 links per page and many times have more than 200 links on each page due to links and images in the header, footer, guided navigation and then the body product grid and content. When I use most on-page link counting tools like SEO x-ray and the SEO Moz Pro crawl report, I notice that every visible link on the page gets counted. This includes and javascript based links that expand the product grid to 30, 60 or view all, javascript sorting links, javascript links to view customer reviews for each product. etc. There was a QA post here http://www.seomoz.org/q/should-i-nofollow-the-main-navigation-on-certain-pages about nofollowing and page rank sculpting and it seems pretty unanimous that most don't think that page rank sculpting is very valuable. So my question is, are the javascript links on pages that don't link to another page viewed differently by search engines? If so, shouldn't there be a way to see on-page link count minus javascript call links that don't actually link to another page? To expand a bit on my question, we also use nofollow attributes on the text links in the left navigation that are meant for refining products just as the javascript links in the product grid are meant to refine the products, sort them, allow for product comparison, allow for viewing customer reviews, etc. So should it be ok to have 300 links on a page if the unimportant ones that you don't want crawled like the left navigation refinements and product grid javascript links all have rel="nofollow" applied to them? I know that would basicly be PageRank sculting, but it seems like the best options for shopping sites that have a lot of navigation links.
Moz Pro | | abernhardt0