Welcome to the Q&A Forum
Browse the forum for helpful insights and fresh discussions about all things SEO.
But how does Google measure form completions or purchases for rankings?
Again, I'm not talking about Google analytics. We use it heavily for our ecommerce sites. I know how the UA tracking code works. Google claims that they don't use GA data for rankings, and I would tend to believe them.
Thanks for the answer. Spot on.
There's been a lot of speculation on "task completion" and how it relates to ranking. If completing a task is a purchase on an ecommerce site, how is Google measuring it? Is it only through Chrome or by some other means?
how does Google measure when someone completes a form?
is that possible, or is Google just checking to make sure that the cart and the form work correctly? Was that the point of the "Zombie" update?
I guess I should have phrased the question a little differently. This is not related to Google Analytics.
When I do a Google search, Google is able to track my actions, and is probably using the data as a ranking factor. Josh Bachynski did a Whiteboard Friday on it.
https://moz.com/blog/panda-41-google-leaked-dos-and-donts-whiteboard-friday
How is Google able to track user actions after they click on a SERP listing? Where are they getting their data?
Assuming that Google is using click data from users to calculate rankings (bounce rate, time on site, task completion, etc.) where does Google get the data, especially from browsers that aren't Chrome?
We have an ecommerce website, and we show a popup for first time visitors to our desktop site to join our email list. Google has cached pages with the popup.
Can I assume that this is a problem?
Yep. We got hit around 7/22. We're an ecommerce site, and most of our 2000 products have little to no search value. We were focused on roughly 50 products. We took about a 20% hit on desktop and 35% on mobile. We had one high value (100K+ monthly searches) product ranked #1 that dropped to #3, and I wrote the traffic drop off to that. Digging in a little deeper, I realized that we got hit. The traffic dropped maybe 10%, but it's been a slow erosion ever since.
All but a handful of products have descriptions between 40-60 words. We are in the process of creating new content, but it's expensive.
Does it make sense to nofollow product pages that have no search value?
I would agree to a point. At its heart, Google probably uses some form of numerical score for a page as it relates to a query. If a page is a perfect match, it scores 100%. I would also suggest that attaining a perfect score is a virtual impossibility.
The scoring system, however, is dynamic. The page may be perfect for a particular query only at a particular point in time.
These are only a couple of examples.
For practical purposes, the amount of testing, research, etc. to achieve a perfect score at some point delivers diminishing returns. The amount of effort required to push a page from 95% to 100% isn't worth the effort, especially since Google's algorithm is a secret.
Sometimes good is good enough.
If you optimize your content properly, you should be able to rank all of those keywords with 1 page (with great content and links). Those keywords are semantically the same. Hummingbird is a awesome if you know how to leverage it.
One trick is to use parentheses. For example, your opening sentence could be "ImageCo is the world leader in Device Imagery Software (DIS) Solutions."
I would use "DIS Software" and "DIS Solutions" a few times in the document, and I would place those keywords in close proximity to the non-abbreviated keywords so Google links them semantically. I would also use all 3 Keywords in their own H2s.
The trick is not to overdo it and make it look natural. No need to spam it.
I would also be mixing in Proof Terms and Relevant Terms throughout the document with the most important going in the first 2 paragraphs.
For a title tag, I would use:
Device Imagery Software | DIS Solutions | Company Name
It's much easier and cheaper to write 1 page of killer content and build links to it than to build 3.
We are an ecommerce company, and we outsource our mobile site to a service, and our mobile site is m.ourdomain.com. We pass the Google mobile ready test.
Our product page content on the mobile site is woefully thin (typically less than 100 words), and it appears that we got hit with Panda 4.2 on the mobile site. Starting at the end of July, our mobile rankings have dropped, and our mobile traffic is now about half of what it was in July. We are working to correct the content issue but it obviously takes time.
So here's my question - if our mobile site got hit with Panda 4.2, could that have a negative effect on our desktop site?
We are in ecommerce, and there are a few review sites that are dominating the rankings for our products. The sites are very good - very well written content (2000+ words) and visually appealing sites. The 2 main culprits are clearly black hat. One site's backlinks are pure spam, and the other is buying footer and sidebar links.
Will ratting them to Google have any impact? If not, any suggestions on how to compete? Our competing pages are product descriptions, and creating a 2000 word product description seems inappropriate. Also, all of these products are brand new, and due to extensive media spends, the search volume is very high. Since they are beating us to the punch by getting good content posted first, they are proving difficult to displace.
I have Google Alerts set up for our websites. I started receiving alerts every time we add a new product to the websites.
I recently started getting email notifications from Google re: new products on our websites. I am subscribed to Google alerts. Can anyone shed some light on this?
Responding to the original question about crowdsearch.me:
It definitely does work, but the results are a little unpredictable. The best way to use it is to help push pages on page 2 or better up the rankings.
There are other similar services out there, and they are also effective.
Sooner or later the Google boys will figure out how to combat it, but for now, it's a viable black hattish weapon.
I have it on pretty good authority that Google uses secure search for indexing. I've seen some definitive test results that categorically prove that Google uses your browsing activity for finding and indexing pages, plus unofficial confirmation from former Google employees.
It's a dirty little secret.
This would be a great test for Rand Fishkin to build a case study on.
You could display all of it and make some clever use of jump links. Just sayin...
EGOL: Makes perfect sense. This, IMHO, is a bad move by Google. They always say "Create for Humans, not Bots" but proper use of tabbed content does make for better UX. We are both eliminating tabs for rankings. Google as usual talking out of both sides of their mouth. Who wants to spend 5 minutes scrolling to the bottom of a ridiculously long page?
EGOL is the man! We moved some content behind tabs, and our rankings did drop. When we moved it back out, they returned. We had some other issues/changes as well, so I can't 100% vouch for correlation.
One interesting test I did run is to run some searches for sentences that were hidden behind tabs on our site. The tabbed content was found, indicating that it was indexed by Google, so they aren't ignoring it.
We decided to go tabless, and I think we are the better for it, but who knows? If you have an enormous amount of content on a page, I would consider tabs, but I would leave the juiciest bits out in the open. FWIW.
Backlinks from DA 21 sites are nothing to sneeze at. I'll take them all day long. Moz actually just rolled out a great tool called "Spam Score". The lower a spam score the better. It will show you the spam score for any links that Moz has found. Spammy links can affect your rankings, so spammy links should be investigated and possibly disavowed. If your site's own spam score is low, I wouldn't get my panties in a knot.
Sounds like Google's Semantic Search update called Hummingbird. We are seeing a lot of traffic for keywords that don't appear on our pages. Pretty impressive on the part of Big G.
We actually have 2 nearly identical ecommerce websites, with identical product descriptions, and prior to us doing a cross-domain canonical, we would frequently have both sites ranking in the top 20 for a keyword.
The cross domain canonical really helped - massive rankings increase.
The SERP did link to the correct (canonical target) domain. If the canonical tag is on domain1.com/product-a, the SERP was correctly pointed at domain2.com/product-a.
Because the page on Domain 1 is supposed to be de-indexed, I was expecting not to see the page at all. This is my first crack at cross domain canonicals. It's an interesting way for Google to handle it.
BTW, from a rankings perspective, the cross domain canonicals were extremely productive. Domain #2 got some huge rankings increases.
I've been tracking the results closely. I should publish the results when I get a chance. The most important result is that the keywords (+/-700) associated with the canonicals improved by an average of 22 positions over the higher position prior to the canonicals being implemented.
What I mean by that is for a keyword (ex: "widgets"), Domain 1 was Ranked 46, and Domain 2 was ranked 57, our average improvement was to position 24, which is 22 positions better than the higher ranked domain (in this case, Domain 1).
Rankings improvements for keywords already on page 1 or Page 2 increased by an average of 2.5 positions over the better ranked domain.
What was really cool was that when we canonicaled in the "wrong" direction, where the keyword ranked higher on the domain that was getting the canonical tag, the results were indistinguishable from the results where we canonicaled in the "correct" direction.
So, in this case, if a keyword ranked higher on domain1.com, and we canonicaled to domain2.com, the average ranking increases (from the higher ranking position) were almost identical to using canonicals in the "correct" direction (from the lower ranking position).
These are both ecommerce sites with DAs of +/-40.
What was also interesting is that Google accepted the canonicals in cases where our product descriptions were markedly different.
We recently ran cross domain canonicals for 2 of our websites. What's interesting is that when I do a search for ""site:domain1.com "product name"" the Title in the SERPs uses the Domain Name from the site the page has been canonicaled to.
So the title for Domain1 (for the search term above) looks like this: Product Name | Keywords | Domain 2
Interesting quirk. Ha anyone else seen this?
Where this is appearing the most is on cross domain canonicals. We have duplicate content across 2 websites, and we've canonicaled some pages from Site A to Site B, and some from Site B to Site A. In theory, pages that were canonicaled to the other domain should be deindexed. When I run a rankings report, I see pages for the wrong domain ranking, a month later. They are pages with parameters, or old URLs that we've changed. It's like a game of whack a mole. Every time we get a page deindexed, a duplicate with a different parameter takes its place. And this is in spite of calling out these parameters in GWT.
What I imagine is happening is that we have several URLs for the same page indexed. When Google crawls our site, it is correctly canonicaling the page it crawls. In the rankings, however, Google is probably pulling a duplicate page out of its index, and ranking it without crawling it. If it was crawling it, Google would see the canonical tag, and not rank it. So we have an ongoing battle to get Google to crawl the page it just pulled out of its index to see the the canonical tag.
The reason for all this is that when a page cross domain canonicals correctly, the rankings for the duplicate page on the other site goes up dramatically. As long as Google keeps ranking the wrong pages, we don't get the rankings bump on the other site.
Due to URL changes and parameters on our ecommerce sites, we have a massive amount of duplicate pages indexed by google, sometimes up to 5 duplicate pages with different URLs.
1. We've instituted canonical tags site wide.
2. We are using the parameters function in Webmaster Tools.
3. We are using 301 redirects on all of the obsolete URLs
4. I have had many of the pages fetched so that Google can see and index the 301s and canonicals.
5. I created HTML sitemaps with the duplicate URLs, and had Google fetch and index the sitemap so that the dupes would get crawled and deindexed.
None of these seems to be terribly effective. Google is indexing pages with parameters in spite of the parameter (clicksource) being called out in GWT. Pages with obsolete URLs are indexed in spite of them having 301 redirects. Google also appears to be ignoring many of our canonical tags as well, despite the pages being identical.
Any ideas on how to clean up the mess?
We have several large eCommerce websites built on Cold Fusion. It is running on IIS, not Apache. We are looking for a blogging package (CMS) that we can bolt on to the website. We don't want the blog residing in a sub-domain. The blog needs to reside in a folder.
NO => blog.mydomain.com
YES => www.mydomain.com/blog/
Has anyone ever adapted Wordpress for this type of situation? Can WordPress reside in a folder? Are there any other suggestions?
Patrick:
I set up the "clicksource" parameter in Google Webmaster Tools in March, and we still have 2400 pages with parameters indexed.
We have several large ecommerce websites, and we've added some tracking parameters to GWT for google to ignore. All pages are correctly canonicaled.
Google has been ignoring the parameters and the canonicals, and still ranks many parametered pages for us. Has anyone run into this?
The website is www.amerimark.com. The number of impressions for "Amerimark" and "Amerimark.com" dropped by 75% starting on March 16 2015. We are ranked #1 for both keywords, and that hasn't changed.
We've been making changes to the site, but this looks more external.
Our flagship website, up until March 16 was getting 1600 impressions and 300 branded clicks per day as per GWT. After 3/16, branded search fell to 300 impressions and 25 clicks per day. Our rankings haven't changed, and neither has our traffic. We would definitely notice the decline in GA and Core Metrics, and it is running about the same.
according to GWT, 75% fewer people started searching for our brand on 3/16, but all of our other metrics are indicating otherwise.
Has anyone seen this before? Is it a tracking issue on our side?
Hi! The spam score for my sites is "--" with no graph. The sites are large ecommerce sites with a ton of branded search and DAs of 50+. Simple question - are these clean spam scores, or has Moz not calculated the scores yet?
I'd also try and group keywords onto pages semantically. The days of every keyword getting its own page are long gone.
Thanks Ryan. The problem is that the URL parameters are created to track clicks from a hero image on our home page. If we block those parameters, we are wasting link juice in a big way.
Several of our websites have experienced a major direct load traffic spike in the last 30 days - roughly 40K new visitors for each site. The bots are emulating IE9 and appear to be hitting our home page and bouncing 100% of the time. The traffic is double our usual volume, or more. Our bounce rates, conversion rate, page views, etc have suffered accordingly. The volume hasn't affected site performance, yet.
Since the traffic is direct load, I can't see this being a negative SEO attack. Plus, our search visibility for everything but our brands is abysmal - there aren't any real rankings to tank.
Our engineers are saying that the IP addresses are diverse, and they aren't seeing any pattern. I also checked GA for traffic locations, and we aren't seeing anything unusual from overseas.It appears that the attack is US based.
Has anyone seen this before?
We have 2 identical ecommerce sites. Using 301 is not an option since both are major brands. We've been testing cross domain canonicals for about 2 dozen products, which were pretty successful. Our rankings generally increased.
Then things got weird. For the most part, canonicaled pages appeared to have passed link juice since the rankings significantly improved on the other site. The clean URLs (www.domain.com/product-name/sku.cfm) disappeared from the rankings, as they are supposed to, but some were replaced by urls with parameters that Google had indexed (apparently duplicate content). ex: (www.domain.com/product-name/sku.cfm?clicksource?3diaftv). The parametered URLs have the correct canonical tags.
In order to try and remove these from Google's index, we:
1. Had the pages fetched in GWT assuming that Google hadn't detected the canonical tage.
2. After we discovered a few hundred of these pages indexed on both sites, we built sitemaps of the offending pages and had the sitemaps fetched.
If anyone has any other ideas, please share.
I'd also check your queries in Google Webmaster tools. You can drill down on them to see impressions by day. If you see that your impressions go to zero for a particular Keyword or keyword stem, it's a sign that you might have gotten a penalty of some kind.
Penguin tends to attack pages. If you see a particular page or pages had significant falloff, it's a bad sign.
We have duplicate eCommerce websites, and we are in the process of implementing cross-domain canonicals. (We can't 301 - both sites are major brands). So far, this is working well - rankings are improving dramatically in most cases.
However, what we are seeing in some cases is that Google has indexed a parameterized page for the site being canonicaled (this is the site that is getting the canonical tag - the "from" page). When this happens, both sites are being ranked, and the parameterized page appears to be blocking the canonical.
The question is, how do I remove canonicaled pages from Google's index? If Google doesn't crawl the page in question, it never sees the canonical tag, and we still have duplicate content. Example:
A. www.domain2.com/productname.cfm%3FclickSource%3DXSELL_PR is ranked at #35, and
B. www.domain1.com/productname.cfm is ranked at #12.
(yes, I know that upper case is bad. We fixed that too.)
Page A has the canonical tag, but page B's rank didn't improve. I know that there are no guarantees that it will improve, but I am seeing a pattern.
Page A appears to be preventing Google from passing link juice via canonical. If Google doesn't crawl Page A, it can't see the rel=canonical tag. We likely have thousands of pages like this. Any ideas? Does it make sense to block the "clicksource" parameter in GWT? That kind of scares me.
Hi! We are in the process of overhauling our websites, and I am hoping that some of you can post URLs for websites that are ranking well and using lots of creative content to help rank their ecommerce category pages. You can post your own, or others that you admire.
We have 2 ecommerce sites that are 95% identical. Both sites carry the same 2000 products, and for the most part, have the identical product descriptions. They both have a lot of branded search, and a considerable amount of domain authority. We are in the process of changing out product descriptions so that they are unique. Certain categories of products rank better on one site than another. When we've deployed unique product descriptions on both sites, we've been able to get some double listings on Page 1 of the SERPs.
The categories on the sites have different names, and our URL structure is www.domain.com/category-name/sub-category-name/product-name.cfm. So even though the product names are the same, the URLs are different including the category names.
We are in the process of flattening our URL structures, eliminating the category and subcategory names from the product URLs: www.domain.com/product-name.cfm.
The upshot is that the product URLs will be the same. Is that going to cause us any ranking issues?
I am not a developer - I am researching this for our team, so please, be gentle... I am also not quite sure how to ask this question.
We want to serve up custom pages for visitors from Google organic. We aren't doing anything underhanded - the pages will have very small differences that will not affect our rankings and won't land us in Google jail.
When a Google visitor hits one of our pages, what specific piece of data are we looking for to determine:
a. It's a Google visitor
b. He/she came from organic results.
I need to tell our developers to look for something that triggers the custom page. It's the same data that Google Analytics uses to trigger the appropriate visitor type.
Please pardon my naivete.
Specifically, how do we determine the medium and source of a visitor?
If we want to change the UX for different types of visitors (ex: direct, google PPC, google organic, bing organic, etc.) what is the best way to do it? With Universal Analytics, where can we find the source and medium?
Without Google Analytics, how would we be able to identify visitors who are from Google Organic? What's the cookie?
Brady:
Thanks for your advice. We are de-indexing as a test to see if our rankings are somehow being constrained because of duplicate content. Our rankings are not reacting as they should to our link building efforts, and I believe that duplicate content is the issue. With this test, I am trying to understand how Google sees and connects these 2 sites. If Google does connect the sites, then we must canonical since Google won't let us have 2 slots on page 1 for the same KW. If Google doesn't connect the sites, then we can theoretically get 2 listings on page 1 if our content is unique.
My hypothesis is that our massive duplicate content is having a negative impact on rankings. Google might be hitting us with a minor Panda slap, or the competing content is somehow hurting us algorithmicly. If we deindex a competing page, if we are being hurt by the algo, the remaining page should make a bump up.
I am pretty certain that canonicals will have a positive impact on our rankings. The question I am testing for is "do we have to canonical"? If we don't, then we have a decision to make - do we try to rank both sites for a KW, or canonical and focus on 1.
Does anyone have any experience with this situation?
We have 2 ecommerce websites that carry 90% of the same products, with mostly duplicate product descriptions across domains. We will be running some tests shortly.
Question 1:
If we deindex a group of product pages on Site A, should we see an increase in ranking for the same products on Site B? I know nothing is certain, just curious to hear your input.
The same 2 domains have different niche authorities. One is healthcare products, the other is general merchandise. We've seen this because different products rank higher on 1 domain or the other. Both sites have the same Moz Domain Authority (42, go figure). We are strongly considering cross domain canonicals.
Question 2
Does niche authority transfer with a cross domain canonical? In other words, for a particular product, will it rank the same on both domains regardless of which direction we canonical? Ex:
Site A: Healthcare Products, Site B: General Merchandise. I have a health product that ranks #15 on site A, and #30 on site B. If I use rel=canonical for this product on site B pointing at the same product on Site A, will the ranking be the same if I use Rel=canonical from Site A to Site B? Again, best guess is fine.
Question 3:
These domains have similar category page structures, URLs, etc, but feature different products for a particular category. Since the pages are different, will cross domain canonicals be honored by Google?
IMHO, Google has essentially tied our sites together, looking at the rankings and other metrics. We've upgraded some product pages with the same result - the upgraded page and the old page on the other side are stuck on page 2. They even mirror each other in the SERPs. They both move up and down by similar numbers. If Page A drops 2 slots, page B drops 2 slots. It's like the pages are attached at the hip...
This discussion is strictly theoretical... I won't hold anyone to their answer.
If I have 2 websites that are identical in every way and let's say the domain authority for both is 40, and I 301 redirect one site to the other, what would the DA become?
Same question for single pages, both with a PA of 40. If I 301 redirect one page to the other, what does the PA become for the remaining page?
It's a big site, and the most important category page. It explains a lot.
We are a large catalog company with thousands of products across 2 different domains. Google clearly knows that the sites are connected. Both domains are fairly well known brands - thousands of branded searches for each site per month. Roughly half of our products overlap - they appear on both sites. We have a known duplicate content issue - both sites having exactly the same product descriptions, and we are working on it.
We've seen that when a product has different content on the 2 sites, frequently, both pages get to page 2 of the SERPs, but that's as far as it goes, despite aggressive white hat link building tactics.
1. Is it possible to get the same product pages on page 1 of the SERPs for both sites? (I think I know the answer...)
2. Should we be canonicalizing (is that a word?) products across the sites? This would get tricky - both sites have roughly the same domain authority, but in different niches. Certain products and keywords naturally rank better on 1 site or the other depending on the niche.
We are running an A/B split test (Started on 12/12), and a few days after we started the test, we fell from position 9/10 (about 3 weeks on page 1) to position 11/13, and we've been there ever since.
We are still running the test. We disallowed the test page in robots.txt, but that's all.
Theoretically, Google could crawl the site, and find the page, but then the page is disallowed. The alternate page is not indexed.
Could this explain the rankings drop?