This case study was created using an LRT Superhero account.
Some of the use cases explained in this case study are not available in lower plans.
The LRT Superhero Plan (and higher) includes all our 25 link data sources and allows you to perform link risk management, competitive research, professional SEO and backlink analysis for your own or your competitor's sites. You get to see your website's full backlink profile picture and this can make all the difference for your SEO success.
Did Google Deliberately Penalize a Nest Competitor?
In January 2014, vivint.com's SEO visibility plummeted. Was it a common penalty or something sinister? Some accused Google of assassinating the site, putting Matt Cutts on the defensive. After all, Vivint specializes in home automation, just like a recent Google acquisition, Nest.
Is Google really assassinating sites' search rankings to boost its own products? Many people are quick to believe it. After all, Google isn't exactly transparent.
But the real story is far more complex. Getting to the truth is going to take us down many different paths. Fortunately for us, Rick Lomas is an expert guide of the back alleys and hidden passageways of SEO. You may be surprised where he leads us.
We look forward to your feedback and always appreciate you sharing the work of our Certified LRT Professionals and Xperts.
Enjoy & Learn
Christoph C. Cemper
Table of contents
- 1. About this case study
- 2. What happened in May 2014 (featuring Matt Cutts)
- 3. Who is Nest Labs?
- 4. Who is Vivint?
- 5.0 Is Vivint a competitor to Nest?
- 6.0 Was Vivint violating Google's Guidelines?
- 6.1 Links on anadesign.info
- 6.2 Links on womenspk.com
- 6.3 Links on frugalful.com
- 6.4 Links on doyoulovewhereyoulive.com
- 6.5 Links on arch.itect.us
- 6.6 Freshome.com and calling BS on Matt's response
- 6.7 Vivint had 25,000+ links from a site with a paid relationship
- 7.0 Link Detox on Vivint.com
- 8.0 Link Detox on Nest.com
- 9.0 CDTOX on Vivint, Nest and some similar companies
- 10 Update: June 26, 2014
- 11 Conclusions
1. About this case study
This is not the usual type of case study in which we talk about why a site lost its traffic or how it regained its traffic. It is true that Vivint (vivint.com) lost their traffic in January 2014 and then regained some of it in March 2014 due to a Google manual spam action being revoked, but this time there is more to it. It concerns the timing of their manual penalty and accusations that Google deliberately singled out Vivint for their own gain.
2. What happened in May 2014 (featuring Matt Cutts)
2.1 The first Pando article: May 29, 2014
An article was published May 29, 2014 on Pando.com titled,
"After Google bought Nest, it removed one of the company’s biggest competitors from search results."
The article was written by James Robinson. You can read it for yourself here:
In the article, Robinson states:
"In the middle of January, Vivint, the Utah-based home automation company that also produces smart thermostats, found itself with a surprising new rival. Google bought Nest and by virtue of acquisition Vivint was suddenly competing head to head with the Silicon Valley search giant.
But Vivint — which was purchased by Blackstone in 2012 — certainly didn’t expect what happened next. Just two weeks later, Vivint was delisted from Google’s search results."
In the article, James Robinson decided Vivint was a competitor to Nest. The article accuses Google of taking manual action on Vivint because they are a competitor. Nest Labs was not cheap; it cost Google $3.2 billion, which is a staggering amount of money for a company that only really has two products – a thermostat and a smoke alarm. Nest was purchased by Google who announced it on January 13, 2014.
2.2 Matt Cutts responds to the Pando article: May 29, 2014
Matt Cutts, head of Google Webspam responded to the Pando article that evening, not on Pando, but on Hacker News, which is part of Y Combinator. Matt is no stranger on Y Combinator and has currently made nine submissions and tons of comments. You can read Matt's reply here:
Matt Cutts said,
"It's a shame that Pando's inquiry didn't make it to me, because the suggestion that Google took action on vivint.com because it was somehow related to Nest is silly. As part of a crackdown on a spammy blog posting network, we took action on vivint.com – along with hundreds of other sites at the same time that were attempting to spam search results.
We took action on vivint.com because it was spamming with low-quality or spam articles like
and a bunch more links, not to mention 25,000+ links from a site with a paid relationship where the links should have been nofollowed.
When we took webspam action, we alerted Vivint via a notice in Webmaster Tools about unnatural links to their site. And when Vivint had done sufficient work to clean up the spammy links, we granted their reconsideration request. This had nothing whatsoever to do with Nest. The webspam team caught Vivint spamming. We held them (along with many other sites using the same spammy guest post network) accountable until they cleaned the spam up. That's all."
2.3 Pando's response to Matt Cutts: May 30, 2014
Robinson responded to the Matt Cutts post on Pando with a second article titled,
“Google’s webspam head says Pando’s reporting "silly," wants you to know Google is totally transparent.”
It was clearly done in a hurry as the grammar and punctuation in the title really was that poor!
Robinson defended himself in the article:
"Cutts’ assessment of Vivint’s wrongdoing is exactly what we described in our article — no one is disputing that Vivint violated Google’s search rules. We confirmed that wrongdoing with the help of Search Engine Marketing (SEM) expert Mike Templeman of Foxtail Marketing who is quoted at length in our reporting.
For anyone not familiar with SEM-parlance, links coded as “nofollow” don’t show up on Google’s map. All advertising — such as banner ads — should be coded as “nofollow” to avoid being classified as a paid link. As Templeman explained, in clearly sloppy fashion someone at Vivint had not correctly coded the company’s banner advertising, making it appear distinctly spammy."
2.4 The Hacker News thread on Y Combinator
Meanwhile, the Hacker News thread over on Y Combinator built up very quickly. Matt Cutts responded several times to comments on the thread, and the response was somewhat lively to say the least. I'll pick out some points from the thread along the way, but for the moment, we have some interesting things to investigate. Ultimately we will need to answer these questions:
- Is Vivint really a competitor to Nest?
- Was Vivint really violating Google Guidelines?
- Is Nest violating Google Guidelines at the moment?
- Was Pando's Google accusation fair?
- Did Google take action on Vivint because they were a competitor to Nest?
3. Who is Nest Labs?
"Nest Labs is a home automation company headquartered in Palo Alto, California, that designs and manufactures sensor-driven, Wi-Fi-enabled, self-learning, programmable thermostats and smoke detectors. Co-founded by former Apple engineers Tony Fadell and Matt Rogers in 2010, the start-up company quickly grew to have more than 130 employees by the end of 2012.
The company introduced its first product, the Nest Learning Thermostat, in 2011. In October 2013, Nest Labs announced the Nest Protect smoke and carbon monoxide detector."
The Wikipedia entry also states that in 2013 Nest Labs had 200 employees.
3.1 Nest products
At the time of writing, Nest only has two products:
3.1.1 Nest Thermostat
Nest Thermostat is a self-learning smart thermostat. Nest says, "Most people leave the house at one temperature and forget to change it. So the Nest Learning Thermostat learns your schedule, programs itself and can be controlled from your phone. Teach it well and the Nest Thermostat can lower your heating and cooling bills up to 20%."
3.1.2 Nest Protect
Nest Protect is a smoke and carbon monoxide alarm, which talks to you! Nest says,
"It’s meant to be piercing, unavoidable, scary. But we hear smoke alarms cry wolf so often that the alarm isn’t doing its job anymore. When it goes off, we’re not scared—we’re annoyed. And we take the batteries out.
Because it’s just a sound. An irritating sound that could mean there’s an emergency or it could mean you’ve burned the popcorn. So Nest Protect does more than alarm you. It talks to you."
It doesn't sound much for $3.2 billion, does it? But, of course, Google isn’t really interested in keeping you warm and safe, but they are interested in moving towards 'The Connected Home,” and (more importantly) finding out what you do when you are at home. Scary? Maybe. There is a good PCWorld article here that you might find interesting:
3.2 Nest SEO Visibility
Nest has nothing at all to worry about; they have a lovely upwards trend.
3.3. Nest Ranking Keywords
Nest Labs has the #1 spot for “thermostat,” which is where they should be if they make the world's best thermostat. They also rank #1 for “nest,” in the same way that Apple.com ranks for “apple.” In both of these cases, only the second Wikipedia entries refer to the structure made by birds and the fruit respectively. It seems odd these days that the technology companies take preference over the natural objects that have been in existence for millions of years! Nest Labs has an interesting and powerful connection with National Grid, too. Nest says,
"National Grid knows the Nest Learning Thermostat will help you save energy. So they’ve partnered with Nest in Massachusetts and Rhode Island to give you a $100 instant rebate when you buy online."
A consequence of this is that they rank on the first page of Google for “national grid.” This is an interesting collection of keyphrases that seem to be more aimed at building brand awareness, rather than getting sales.
4. Who is Vivint?
"Vivint, Inc. (formerly known as APX Alarm Security Solutions Inc.) is a privately held company providing home security, home automation and energy conservation services in the United States, Canada and New Zealand. Vivint Inc. is headquartered in Provo, Utah. Vivint serves more than 750,000 installed systems across the United States, Canada and New Zealand, and retains over 7,000 employees."
The Wikipedia entry also says,
"Prosecutors in at least eleven different states have sanctioned Vivint for deceptive sales practices including high pressure door-to-door sales tactics and other violations."
I was quite shocked at reading that last part, but a comment by 'sivetic' on the Hacker thread echoes this:
In late 2010, the company launched a company-wide rebrand to better represent the potential and future direction of the company. On February 1, 2011, APX Alarm changed its name to Vivint. It’s based on the words "vive," which means "to live," and "intelligent." Vivint says that they are dedicated to helping their customers live intelligently by providing simple technology for a smarter life. In June 2011, Vivint's Home Automation and Advanced Security packages received a Consumers Digest "Best Buy" rating.
4.1 Vivint Products
Vivint was predominantly a home security company, but since 2011 has been moving more into the home automation niche. They also provide energy solutions such as solar panels.
4.2 Vivint SEO Visibility
Using Searchmetrics we can see a huge drop in visibility around January 19, 2014 and recovery in March 2014. We can safely say that this corresponds to the Matt Cutts thread on Hacker News:
"When we took webspam action, we alerted Vivint via a notice in Webmaster Tools about unnatural links to their site. And when Vivint had done sufficient work to clean up the spammy links, we granted their reconsideration request."
Unfortunately for Vivint, their revoking of the manual spam action seems to have had some effect, but probably not quite what they were hoping for. With any Google manual spam action for unnatural inbound links, prevention with Link Risk Management is always better than a cure.
4.3 Vivint Ranking Keyphrases
Even though Vivint has had their manual penalty revoked, they are still struggling with some keyphrases, in particular “home automation” is only at #22. The third page of Google is a dismal place for anyone to be. It's not all doom and gloom though, they do have some good phrases on the first page. For their old brand they have a #3 spot for “apx alarm,” and they have the phrases “home security systems,” “home security” and “home security system” all on the first page.
5.0 Is Vivint a competitor to Nest?
Here I would disagree with the Pando article and say that they are not. The structure of the companies is completely different. Nest is all about making the best and most advanced products that can be used for home automation, whereas Vivint probably still makes the bulk of their money installing home security systems. I can't imagine a team of Nest/Google salesmen knocking on the doors of home owners trying to sell them a $249 thermostat. The two sites don't even rank for one similar keyphrase. Even though their niches are related, it is important that Vivint ranks for “home security” as much as Nest ranks for “thermostat.”
I actually think that Google's acquisition of Nest has more to do with them being in competition with Apple. There's an interesting video about this here on Bloomberg.com:
Nest was founded by Tony Fadell in May 2010, but before that he worked at Apple. He served as the Senior Vice President of the iPod Division at Apple Inc. from March 2006 to November 2008. He is known as "one of the fathers of the iPod" for his work on the first generations of Apple's music player.
If you have any thoughts about Google v. Apple or “The Connected Home,” please leave a comment below.
6.0 Was Vivint violating Google's Guidelines?
We have the answer direct from Matt Cutts: YES! But some people in the Hacker thread were not happy with what Matt said, so let's look at that in more detail:
6.1 Links on anadesign.info
Matt Cutts said,
We took action on vivint.com because it was spamming with low-quality or spam articles like
The website anadesign.info no longer exists, which is why Matt pulled the example from archive.org:
This is a DoFollow link with the anchor text “Albuquerque NM home security system” linking to www.vivint.com/en/city/nm/albuquerque. Anchor text links like this don't happen naturally and violate Google's guidelines.
6.1.1 More links like anadesign.info
To find some more like this we just need to go to www.vivint.com/en/city/nm/albuquerque and run a Quick Backlinks or QBL report. The easiest way to do do this is by having a [QBL Page] bookmarklet on your browser’s bookmark toolbar. Just go to the page you want to see the backlinks for, click the bookmarklet and a QBL will start.
This QBL report actually only finds six links, but let's look at them anyway:
The first one, anadesign.info, we know about already. The fifth one, abqhomeexp.com, looks quite legitimate on a page about “Exhibitors who were at the March 2014 EXPO” it's not a particularly good link as it is on a page with a load of other links, but I can see why it is there. The other ones are NoFollow links.
Paperblog.com has an article about preparing for your new baby.
One of the five fun ways you can prepare for the birth of your child is by installing a home security system! Yes really...
This has two links to Vivint on this page, they are both NoFollow, but I have my suspicions that they were not always so. Unfortunately, this page is not available in archive.org, so I can't check it. But does this look natural to you? No.
Giveaways4mom is a guest blogging site which has an audience of which 90% are mothers. Here is a post called “Technology for Moms: Because Sometimes We Just Can’t Do It All.” These mothers must really love to read about things they can't possibly have time to do:
"Most moms know that using a programmable thermostat is a great way to cut down on energy consumption, but who has time to program it and make changes as necessary?"
The article ends in the usual resource box style with an anchor text money link:
The links on giveaways4mom.com are NoFollow too, but whilst looking around the site I found that, on February 13, 2014, there was an article posted about WordPress plugins:
Regarding plugins, they mention “Auto No Follow Links,” saying, "Auto No Follow Links is a life savior for review and giveaway blogs. Affiliate links and sponsored links should always be no follow and this plugin does the job." This makes me wonder if they had run into a few problems with DoFollow links.
This is an article about using your smartphone for various things like home automation. It follows a similar kind of format as giveaways4mom.com with a resource box at the end of the post with an anchor text Money link:
Once again, this is a NoFollow link. Although it shouldn't cause a problem, the link is certainly unnatural.
6.2 Links on womenspk.com
Matt Cutts also mentioned this link on the Hacker thread:
Here there are two DoFollow links, one with a compound anchor text of “Vivint Home Security” linking to www.vivint.com/en/city/nc/wilmington, and the other with the anchor text “Ann Arbor MI” linking to www.vivint.com/en/city/mi/ann-arbor. For a moment I thought Ann Arbor was a person, but it’s actually the fifth largest city in Michigan, USA.
6.2.1 Doorway Pages
The pages that these links are linking to are all pretty similar, a basic landing page with just the locations changed. This might not just be a Penguin issue but also Panda one, too. I'm guessing that pre-penalty, Vivint was probably ranking #1 for “home security ann arbor,” but now they are not in the first 100 results on Google.com.
To find out more about these doorway pages I took a guess at the sitemap URL: http://www.vivint.com/sitemap.xml - Boom! Got it ☺. I copied the contents of this page into Excel, did some sorting, and then just made a list of the doorway pages. These are all the ones that begin with www.vivint.com/en/city - there are 1,819 of these from:
All the pages are pretty much identical. Every state is included, even Hawaii:
I can't resist digging into these doorway pages a little more.
6.2.2 Using the Juice tool on Vivint's doorway pages
For a bit of fun I thought I would put these 1,819 URLs into the Juice Tool. I just copied the URLs into the “URLs to Analyze” section and also selected “Detail Analysis” with some metrics enabled. I was particularly interested in how these doorway pages were being linked.
The results were pretty much as I suspected. The Power*Trust value of the doorway pages only varies between 0 and 2, with 1 being most popular.
I decided to dig a bit further from the Juice Tool report and show you some more bad links that Vivint has, as well as some they have removed. I made a short video about this here:
In the video you can see that Vivint built their own backlinks to their doorway pages, concentrating on the high end cites like Providence, Rhode Island and Las Vegas, Nevada. However, it does look like they have certainly made some effort to remove these links. By filtering the data table from the Juice report for BL > 0, I can see that 412 of these doorway pages have links to them. On the assumption that there is no content value in these pages, there is no reason for anyone to link naturally to them. When I looked further at some of these pages many had the same style of spammy blog post linking to them, so we can assume that all the linking was done artificially by Vivint.
6.2.3 Do the doorway pages work?
For low competition keyphrases like “home security north pole,” Vivint is enjoying a #1 spot with the URL:
I'm not sure how much business they will get in this remote part of Alaska, but with virtually no competition they will probably be able to keep their position for a while with no further effort.
For high competition keyphrases like “home security los angeles,” Vivint only makes it to the 12th spot underneath the sponsored and local results for this URL:
Los Angeles is the most populated city in California and the second most populated city in the United States. Although crime in Los Angeles is at a 50 year low at the moment, the city is home to 45,000 gang members, organized into 450 gangs. Home security is serious business in Los Angeles. The economy of Los Angeles is driven by international trade, entertainment, aerospace, technology, petroleum, fashion, apparel, and tourism. Los Angeles is also the largest manufacturing center in the western United States. Combine this with the fact that Los Angeles is the third largest economic center in the world, you can see that this is one of the wealthiest places on Earth. The fact that Vivint is not ranking on the first page of Google for “home security los angeles” must be costing them a fortune. How hard would it be to make a proper page about the services they offer in Los Angeles and its suburbs?
If I was responsible for the Vivint web site, I would watch this Matt Cutts video about “Thin content with little or no added value”:
6.3 Links on frugalful.com
Matt Cutts also mentioned this link on the Hacker thread:
This is an article called “Five Surprising Ways To Save Money On Your Homes Expenses.” Anybody who reads this nonsense really needs to get a life:
"Instead of buying brand-name cleaning products, make your own by using baking soda, water, and vinegar mixtures that will work just as effectively on the floors, countertops, and toilets."
To save even more money, install a home security monitoring system. Presumably you will be able to do this with the money you save by making your own cleaning products.
In this case they have now removed the links, but for some reason made them a different color. I don't know why, but it helped me find them! There is now one of those annoying infolink popups that comes up when you hover over “home automation.” These were presumably anchor text links of “Vivint home automation” and “home security monitoring.” Vivint has done the right thing by removing the links. The article content looked like it was from a Fiverr gig, nothing more.
6.4 Links on doyoulovewhereyoulive.com
Matt Cutts also mentioned this link on the Hacker thread:
As usual, this has the standard unnatural anchor text:
This is what Matt Cutts was objecting to, but there is something else here. At the end of the article it says,
"You're reading Top 10 Benefits of Automating Your Home originally posted on Freshome.
The post Top 10 Benefits of Automating Your Home appeared first on Freshome.com."
We will take a look at Freshome in a minute, but first let’s look at the final Matt Cutts sample link.
6.5 Links on arch.itect.us
Matt Cutts also mentioned this link on the Hacker thread:
I would look at this, but the site is broken. I tried looking on archive.org and there is a sample from May 30, 2014, but it appeared to be broken then as well. Luckily, this site was mentioned in the Hacker News thread by someone reacting aggressively to the comments by Matt Cutts. I'll talk about that in the next section.
6.6 Freshome.com and calling BS on Matt's response
This is getting juicy now. Enter, Theo_Stirling;
To answer Theo_Stirling's first question, I needed to take a look at this link:
The first thing that struck me about freshome.com was that the site was actually interesting, unlike the others I had looked at so far. I like architecture, I like houses, and I always dream of living somewhere hot and sunny with a sea view. There were plenty of homes like that on freshome.com. So where was our link to Vivint? Gone, it had been removed. Using archive.org I was able to get a snapshot from May 11, 2013 which showed the link still in place;
Now we can answer Theo_Stirling's first question:
Q. Should sites/brands be worried about content that gets scraped and put on low quality sites?
A. Yes, they should, especially if they are using Money keyphrase anchor text in their original article which they later change their mind about!
This scenario will be a familiar one for all those who have published press releases with heavily overdone anchor text. Although this was a popular thing to do before 2012, it is now suicide. Once the press release gets syndicated around the internet, all that overdone anchor text will be nearly impossible to change or remove.
Theo_Stirling's second question was,
Q. What about Nest posting on Freshome.com as well found here:
THEN we see the scrapes on the same sites found here:
Notice how both NEST articles are found on doyoulovewhereyoulive.com AND arch.itect.us?
Case in point, Nest has links from the EXACT SAME SITES and didn't get penalized. This raises a huge red flag. If these links are the reason that Vivint was penalized, I'd love to hear how Nest somehow managed to escape a penalty.
A. There is no red flag raised at all. Nobody is going to get penalized for a couple of bad links on doyoulovewhereyoulive.com or arch.itect.us. These were just ones that Matt Cutts chose as examples. As I pointed out in section 6.1.1 of this article, you can easily use the sample links provided by Matt Cutts to find similar examples and that is what it is all about - does your backlink profile look natural? You might also note that even back in pre-Penguin 2011, Nest just linked their brand and nothing else, no keyword, just “Nest.” We know now that brand linking anchors are generally a good thing, but who knew that in 2011? As a consequence of this, the brand links that Nest has on these sites is doing them no harm at all.
To finish this answer off properly, we will need to compare Nest against Vivint with the Competitive Link Detox Tool. But first we need to investigate the last point made by Matt Cutts.
6.7 Vivint had 25,000+ links from a site with a paid relationship
As well as the five sample dodgy links that Matt Cutts revealed, at the end of his initial post he also said,
"....not to mention 25,000+ links from a site with a paid relationship where the links should have been nofollowed.”
We can use the Backlink Profiler (BLP) to investigate this further.
6.7.1 BLP Settings for Vivint
To find out more about the alleged 25,000+ links to Vivint, I used the Back Link Profiler with the sitewide filter turned off. Remember that Vivint has now cleaned up their link profile enough to get out of a manual penalty. This means that some links may now be gone, but we would like to see them in this case. For this reason we are also going to make sure that we include the dropped links:
This took a long while to run; I'd never seen a BLP take so long. There were a lot of links to find!
6.7.2 The BLP Results
I was right about the number of links:
This looks promising as I have 36,957 links from only 667 domains, so we are looking at some serious sitewide links in there. The anchor text cloud, pie chart and statistics show this very clearly:
What we are seeing here is that there are 30,784 image links with no alt-text and 13,652 image links with alt-text = vivint. This makes up an astonishing 95.6% of the total profile.
To find the offending links I did some investigating by looking at the data table of the BLP report. I made a video about doing this:
6.7.3 The Vivint sitewide link on byucougars.com
I think I have identified this link as the one that Matt Cutts mentioned:
This link is NoFollow, but we can look back a few months in archive.org and see the same page from January 2014:
The HTML code for this link is:
<a href="/web/20140108144744/http://www.vivint.com/" target="_blank"><img src="/web/20140108144744im_/http://byucougars.com/sites/all/modules/custom/premiere_partners/images/vivintPartners.jpg" alt=""></a>
This is a DoFollow link that has been changed to NoFollow since January 2014.
6.7.4 What is byucougars.com?
The website byucougars.com is the official home of the BYU Cougars. The BYU Cougars are the collegiate athletic teams that represent Brigham Young University. Brigham Young is a private university located in Provo, Utah. It is owned and operated by The Church of Jesus Christ of Latter-Day Saints and, excluding online students, is the largest religious university and one of the largest private universities in the U.S.
Vivint is also based in Provo, Utah, too. If you look at this page about the Brigham Young Premiere Corporate Partners:
you can clearly see the paid relationship that Vivint has with BYU. But more fun than that, I like these comments on the message board, cougarboard.com. Their relationship with BYU didn’t seem too popular back in 2012:
6.7.5 How easy is it to get a link at byucougars.com?
For a bit of fun I found the page about sponsoring BYU and got the email address to offer sponsorship. I wrote them an email:
I used a false name that I use sometimes if I want to be disconnected from the SEO world. Amazingly enough, after eight days I have received no reply. I am guessing it probably helps if you went to BYU and live in Provo, Utah.
This article on Forbes.com talks about Todd Pederson, founder of Vivint;
The article states,
"Todd ... counts Brigham Young University as his alma mater—no, wait a minute—he actually didn’t complete his BYU education. Like fellow Utah entrepreneur Josh James, CEO and Founder of DOMO (one of last year’s Most Promising Companies), he dropped out of BYU to continue building the startup company that the world now knows as Vivint."
7.0 Link Detox on Vivint.com
Obviously, without the co-operation of Vivint we do not have access to Google Webmaster Tools, so we do not have any extra links or their disavow file to upload into Link Detox.
If Vivint has not done a lot of work cleaning up their link profile, they would not have got their manual penalty revoked, so I want to look at their dropped links. You can deselect “remove dropped links” in the advanced settings.
In June 2014, the new option of choosing to take into account NoFollow links or not was introduced.
The new functionality allows the evaluation of all links and ignores all links with NoFollow.
This means that, in this mode, the Link Detox Risk (DTOXRISK™) rating is ZERO for NoFollow links.
This is helpful because the first priority in proofing and cleaning up bad links should always be to start with DoFollow links. Personally, I’m not convinced that NoFollow links are ignored by Google, so I am going to take them into account. So the top of the Link Detox looks like this:
7.1. Link Risk of Vivint.com measured by DTOXRISK™
Link Detox found 203,989 links, but then filtered 201,845 links from the report because of excessive sitewide links.
Next, I spent a while classifying the keyphrases into “Brand,” “Compound,” “Money” and “Other.” The important thing here is to be consistent with any decisions I made when classifying the keywords of Vivint. After reprocessing the report, the Average Link Detox Risk score (DTOXRISK™) is 1082, a high DTOXRISK™. Normally a site like this with a DTOXRISK > 1000 is very likely to trigger a Penguin penalty.
Do remember that this report also includes dropped links, so the risk displayed here will not be a true representation of how the site is now. For this reason, I like to use a technique that I haven’t seen anyone else do which I find very helpful.
7.1.1 Creating a “Dropped Links” slice and a “No Dropped Links” slice.
I created a video to explain how I do this:
After creating the “No Dropped Links” slice, the Vivint link risk is much less:
Vivint must have done a pretty good job removing links to get their manual penalty lifted. This is quite an acceptable link risk for most niches.
7.2 Dropped Link Analysis of Vivint.com
Looking at the “Dropped Links” slice I made for Vivint, we can see that the link risk is quite scary. It is good these links are now gone:
There are 1,611 pages in the Link Detox which previously had a link to vivint.com, but no longer do. So what sort of links were these? The easiest way to find some examples is to use the Link Detox Screener on the results from the “Dropped Links” slice I made in the previous step. As I’m looking for a particularly bad example, I filtered the table so I am looking at only the links with a High Link Audit Priority and a High DTOXRISK™ risk. This gives 146 pages to look at. By page 11, I found a good example:
The link is now gone, of course, but I can guess where it was. There is a paragraph titled, “Home Security,” so I am pretty sure that is where it was. By putting the link into archive.org I found this result from October 18, 2013:
There are actually two links here, both going to the Vivint doorway pages for Granite City, IL and Rapid City, SD. This particular link triggers 4 rules in Link Detox:
- TOX3 - The Link Detox Genesis™ algorithm classified this link as highly unnatural.
- SUSP4 - Domain's homepage has CEMPER TitleRank-home™ 30+, CEMPER Power*Trust™ Domain < 5 and CEMPER Power*Trust™ < 5.
- SUSP1 - Page has no CEMPER Power*Trust™ and CEMPER Power*Trust™ Domain < 5 - a page without external links on a weak domain.
- SUSP19 - Old domain with no homepage PageRank™.
It’s a similar pattern to the type of links that we have seen already, and this certainly did not happen naturally. This is getting very easy to spot now, just as Matt Cutts said in the Hacker News thread. All I need to do is to look for any title like “X ways to improve something in your home” and then look for two links to Vivint doorway pages. As Vivint has most of its business in the US, I was intrigued to see a UK blog post:
The links are now gone, but the last paragraph mentions Vivint:
“You can also receive notice remotely on your smartphone. Home security companies such as Vivint home automation can monitor the sensor 24/7.”
I am guessing this is where we will find the links if we look back in time. Archive.org has a snapshot from January 26, 2013:
These links go to the Vivint doorway pages for Greely, CO and Temecula, CA. This is ridiculous for a UK blog about home improvements.
7.3 Really Bad Links to Vivint.com
To judge Vivint and Nest Labs equally, I took my slice idea one step further and created the “Really Bad Links” slice. I used these filters on the data table:
- Anchor Text ≠ [LinkNotFound] - as I did before
- Classification = M - we are interested in money keyword anchor text
- Link Status = FOLLOW - this is an extra metric you have to enable, just above the table
- Link Audit Priority = High - what else?
- Link Detox Risk (DTOXRISK™) > 1000 - this is really high!
- Rules ≈ "TOX1" - Google doesn't index the pages that have these links, so what does that tell you?
The result was just three really bad links from two domains:
The first link is on a domain which has expired, so no threat there. The second and third examples are the usual random blog posts with a link to a Vivint doorway page:
Either they got away with this one, or they disavowed it. I could carry on all day finding examples like this, but it is probably time to look at the Link Detox of nest.com.
8.0 Link Detox on Nest.com
Without the co-operation of Nest we do not have access to Google Webmaster Tools, so we do not have any extra links or any disavow file to upload into Link Detox.
I want to make some close comparisons between Vivint and Nest, so I am also going to look at Nest’s dropped links:
8.1 Link Risk of Nest.com
Link Detox found 151,178 links, but then filtered 125,396 links from the report because of excessive sitewide links.
Next, I spent a while classifying the keyphrases into “Brand,” “Compound,” “Money” and “Other.” The important thing here is to be consistent with any decisions I made when classifying the keywords of Vivint. After reprocessing the report, the Average Link Detox Risk is 400, a moderate risk. I don’t normally see sites with a moderate risk running into Penguin problems, so I think Nest is quite safe here. Do remember that this report also includes dropped links, so I will make a slice as I did with Vivint to see what the Link Detox risk is with no dropped links. The risk is actually stunningly low:
8.2 Dropped Link Analysis of Nest.com
The Link Risk of just the “Dropped Links” slice doesn’t look as bad as I suspected, just 1062 compared to Vivint’s 1936:
There are 5,426 pages in the Link Detox which did have a link to nest.com, but no longer do. This is a lot more than the 1,611 pages that my analysis of Vivint showed. So what sort of links were these? Once again I’m looking for examples of very bad links. I have filtered the table so that I am looking at the links with High Link Audit Priority and a High DTOXRISK™ risk. This gives 327 pages to look at, and I can already see one on page 7. By visiting the page, I can see that the dropped links were actually still there. This can happen if for some reason the page is down or particularly slow when Link Detox is run. This is also why it is good to run Link Detox at regular intervals if you want to maintain your healthy link profile. So let's look at the page:
The example here has two links both with anchor text that I would classify as “other.” The anchor “here” links to http://support.nest.com/customer/portal/articles/182807 and the word “here” links to http://www.nest.com/press/index.html#. This page triggers 3 rules in Link Detox:
- TOX3 - The Link Detox Genesis™ algorithm classified this link as highly unnatural.
- SUSP14 - Page has no PageRank™ but at least some weak links.
- SUSP9 - Domain has the same DNS as other linking domains - possible Link Network.
As the anchor texts on this page are not Money keyphrases, I don’t see this as any major threat.
The next example I found was at http://geosurf.info/tag/nest/.
This is an article about Nest on a low quality site (Power*Trust=2) and there is now no link to nest.com. Even Archive.org doesn’t have any history of this page, so I couldn’t check that. I suspect what might have happened here is that Nest noticed the inbound link, contacted the site owner and asked them to remove the link. As the phrase “more about Business and Nest” needed to link to somewhere, they changed it to http://mashable.com/category/nest/?utm_medium=feed&utm_source=rss. I actually think that a Brand link like this would have been fairly harmless anyway.
I know some readers would love me to find some dirt on Nest at this point, but I honestly can’t. There are quite a lot of links in forums, but even they look genuine. In the forums people are either talking about Google’s acquisition of Nest or the products themselves. I never realized a thermostat would become a “must have” gadget, but it seems to be. Nest seem to be doing their job very well!
8.3 Really Bad Nest Link
Link? Yes, it is only one link! I used the same filtering to create the “Really Bad Links” slice like I did with Vivint:
This is slightly weird in that it is an RSS feed, which makes it a little hard to find the bad link:
We know that the anchor text is “devices,” but that isn't really going to help us find the link. The site is a WordPress blog so I just entered this search string:
This just returns one URL:
This page does feature the word “devices” twice, but in each case it is not a hyperlink. I'm guessing that it has already been removed. This page isn't in Archive.org so I can't check this historically. There is, however, a link to Nest at the bottom of the article:
I'm sorry to disappoint, but this isn't particularly bad! Nest seems to have been very careful about looking after their inbound link profile. I'm sure if I looked hard enough I could find some bad links, but it is certainly not as easy as finding them for Vivint.
9.0 CDTOX on Vivint, Nest and some similar companies
Finally, to get an idea of what the home security/home automation battlefield is like, I am going to do a Competitive Link Detox (CDTOX) comparing Vivint and Nest with some other competitors in the home automation and home security niches:
9.1 Power and Trust in CDTOX
It is interesting to see that when we start looking beyond Nest and Vivint, there are some players with more Power*Trust. In fact, all four of our “outside” competitors have higher Power*Trust, with control4.com and crestron.com doing exceptionally well. These two both have a Power = 4 and Trust >5, having more Trust than Power is always a good thing.
9.2 CDTOX Risk Levels
There were a lot of Referring Root Domains in this report. CDTOX stated, “We filtered from your report: 502,962 links because of excessive sitewide links and 6,287 deleted links.” After its initial processing, I was left with 5,345 anchor texts to classify which took almost a whole day to organize into Brand, Compound, Money and Other. I managed to classify 93% of the links, which is more than sufficient for CDTOX to give accurate results. After classifying the anchor text, the report needs to be reprocessed. The reprocessing of the report took more than one hour, too. There are a lot of links here!
Out of the six sites we compared, Vivint has the highest Domain DTOXRISK™ and Nest has the 2nd lowest. This is almost certainly because Vivint has just got out of a manual penalty and probably has a substantial disavow file in place that we do not know about. It is not often I see the Top 3 Average as low as this, so this does seem to be a niche where you have to be quite careful. As the goal is to blend in with the rest, Vivint is standing out, and Nest is not.
9.3 CDTOX Comparison by Keywords
Once again we can see that Vivint is standing out by having a lower percentage of Brand keywords and the highest percentage of Money keywords when compared with both the Total Average and the Top 3 Average.
9.4 Strong and Healthy Links of Vivint's competitors
One of the great things about CDTOX is the pre-made slice called “Strong Healthy Links of my competitors.” This produces a report of the very best links that your competitors have so you can go after them yourself and crush your competition. Normally I only see a few links here, but in this example there are 192 links like this. The links have Power*Trust values ranging from 10 to 42. Would you like to see them? Sorry, these are too valuable. I'm keeping them to myself unless you want to hire me ☺.
10 Update: June 26, 2014
When I began writing this report at the beginning of June, it was already public knowledge that Vivint had their manual penalty revoked. Matt Cutts stated this on Hacker News on May 29, 2014. We could already see some signs of recovery, but (as we found out earlier) they were still struggling with some keyphrases. Today, things are much better.
10.1 SEO Visibility
This is the kind of Penguin recovery that everyone who has suffered from a manual spam action wants to see. Indeed, their winning keyphrases this week are impressive, too:
Vivint now has results on the first page of Google for “home security systems,” “security systems” and “security system.” As you can see, the search volume and CPC of these phrases is very impressive. There will be some smiling faces this week in Provo, Utah.
10.2 Can we see any changes during June 2014?
There are quite a few different ways to see this, but I thought it was worth looking back to my earliest investigations of Vivint from May 31, 2014 and then comparing them with today. This is a Link Detox report which was one of my reports from when I first looked at Vivint.com. To make this report make sense with the later reports, I reprocessed it. At the same time I re-ran the Link Detox report, creating a new report for June 28, 2014. Now we have a straight comparison of two reports nearly one month apart. There is actually very little change over this period.
May 31, 2014
June 28, 2014
Looking at the links in more detail, there seems to have been very little change during the month of June.
May 31, 2014
June 28, 2014
Vivint has been recovering nicely during June 2014, so the only assumption I can make is that they disavowed a lot of links. As Google recrawls these links, the disavow file is beginning to take effect.
I'll conclude by answering the questions I mentioned earlier (section 2.4). These are my conclusions, formed after writing this case study. If you disagree, please comment below.
11.1 Is Vivint really a competitor to Nest?
No. Although they are both in the home automation niche, they are trying to achieve quite different things. Vivint is interested in selling home automation systems and in the installation of home security systems. Nest has only two products at the moment, but they are interested in making them the best on the market. They don't rank for the same keyphrases, and I don't believe they are trying to rank for the same keyphrases either.
11.2 Was Vivint really violating Google Guidelines?
- I explained their paid association with Brigham Young University which was responsible for more than 25,000 links. This is a violation of Google's Guidelines concerning paid links.
- I found multiple examples of unnatural anchor text which had been used on spammy blogs. This violates Google's Guidelines concerning the manipulation of search results.
- Vivint.com has many “doorway pages” that are almost identical. This violates Google's Quality Guidelines.
11.3 Is Nest violating Google Guidelines at the moment?
No. I examined the link profiles of nest.com and vivint.com in a similar way and I can see no violation of Google's guidelines at all.
11.4 Was Pando's Google accusation fair?
No. The author James Robinson made the assumption that Nest was a direct rival to Vivint without finding out anything about the company and what their goals were.
11.5 Did Google take action on Vivint because they were a competitor to Nest?
No. I know there are plenty of SEOs that would like me to have found out that Google is evil and that Matt Cutts kills kittens for fun, but this isn't the case! Unfortunately for Vivint, this is a fairly typical manual spam action that happened for the correct reasons. I have proved that:
- Vivint had a paid link relationship with sites owned by Brigham Young University. This resulted in thousands of sitewide links.
- Vivint has been building spammy links to hundreds of low quality doorway pages.
This has been an interesting case study of Vivint, which is now a $2 billion enterprise. The manual spam penalty must have cost them a fortune. Vivint's recovery, especially in the month of June 2014, has been very impressive. Vivint has done great work to get their link profile cleaned up. Compared to many sites that have had a serious manual spam penalty, this penalty was revoked fairly quickly. As some of the truly awful links still seem to be in place, it is also proof that disavowing links as well as manually removing links does work. The sad thing, which I see too often, is that this penalty could have been prevented by regular monitoring their backlink profile and disavowing/removing any risky links. If somebody from Vivint is reading this, get in touch with me. I can help. I know your backlink profile very well!
This thorough case study was written by Rick Lomas, Owner of Indexicon, and proud user of LinkResearchTools and Link Detox.
A word from Christoph C. Cemper
Rick demonstrated his expertise by digging into the Vivint.com penalty to determine if they were justly or unjustly penalized. He did a great job uncovering the real story behind the hype. Therefore, I’m very happy to certify Rick Lomas as the latest Certified LRT Xpert by approving and publishing his research on our site.
I look forward to his future work, and personally recommend working with Rick Lomas whenever you get the opportunity!
With the Superhero Plan you can perform link audits, link cleanup, link disavow boost, competitive research, professional SEO and backlink analysis for your own or your competitor's sites.
You can avoid a Google Penguing Penalty! Learn all about the Real-Time Google Penguin Update in this free webinar.
Join our free 21 Day Link Strategy Training below
Signup to receive snack sized bits of knowledge about Link Research Strategies, SEO Tactics & features of LinkResearchTools platform and technologies. Bonus: signup you will also receive a free copy of the eBook "7 Golden Rules of Link Building" for immediate download.