We are tasked with many weird and wonderful jobs throughout the course of a day but this one seemed worthy of a blog post! It is quite well known that we have a team of people who are well versed in all things digital – be it small scale affiliate hobbies we run on the side to managing multi channel projects for our brand clients. This was a job with high stakes – and that always gets us interested.
The client has requested anonymity so I will cover the concepts and activities we undertook to help the client make a decision on the level of risk involved in making an $1,000,000 investment in a business that makes the majority of its income from organic visibility.
What we can say is:
- They are a very large property investment company in the Southern hemisphere with a global presence.
- They own or have investments in a lot of large online property websites across the world including stakes in a large UK brand in this space.
- They are looking to enter a new and fast growing marketplace in South East Asia.
Before they press go on releasing the funds they wanted a risk assessment with regard to potential penalty or down-weighting from Google, in the context of and relative to their peers in the territory represented by the other sites ranking and identified to us.
The job spec
The client wanted answers to some basic but fundamentally important questions about the website and the marketplace in which it operates:
- How stable is their Google visibility?
- How do they stack up vs the rest of the industry?
- Should we buy them at the agreed price (it was 7 figures USD)?
Whilst the report we produced was not the sole driver for the decision making process it did play a critical part in the strategic direction as 80% of the traffic to the current site was via organic search across the globe.
The investigation was limited to historical link-building activity and organic link growth and will not focus on onsite ranking factors and technical SEO practices found on the website.
The methodology used
We were given from the client the URL of the site they were looking to purchase along with 2 more URLs of large portals they considered to be rivals.
- We pulled every piece of link data we could get our hands on.
- We utilised tools like SEMRush to get an overview of the visibility and some historic ranking and search volume data.
- We used LinkRisk in conjunction with Majestic, Ahrefs and OSE to gather and crawl as much link data as we could – We crawled and scored almost 8 million URLs as part of this exercise as we grabbed link data going back 6 years to ensure we had an accurate a picture as we could feasibly get using publicly available data.
- We then manually audited the link profile of the domain the client was looking to purchase.
The link profile of the domain the client was looking to purchase was the smallest of the 3 sites we focussed our efforts on with only 1.76 million linking pages from 12,504 unique referring domains.
The other 2 rivals to give you an idea on scale had 2.9million linked pages from 28,705 domains and 3.08million linked pages from 20,584 domains.
Signals used to analyse the data
We use a combination of signals to help us analyse a link. A lot of what we do is experience based and what looks natural to the passing eye, may not to a trained eye – The trained eye of a link profile auditor – or of a Google spam fighter!
Below are 4 of the many signals we use to evaluate links:
Anchor text as a signal
Anchor text was evaluated against a set of rules to help alleviate any problems that may arise due to the primary “keyword” also being used as a brand term for the main site of interest.
The domain name as a domain.com mention and any supplementary URLs were treated as brand anchors. We were given a master list of industry terms in local and 5 international languages by the client to use as a reference point.
The “Keyword” and “Keyword + Location” will be treated as commercial anchors.
Where it’s difficult to determine whether the anchor is brand or commercial, we will look at the link location to attempt to determine if the link is natural or not.
Variations were translated on a case by case basis to identify the English equivalent to determine whether the anchor is commercial or brand weighted.
When looking for blatant spammy links we also look at the way the anchor text has been written. Capitalisation of the first letter of words for example is a good indication of how natural the link is, and was the case on this project when looking for stand out footer links.
PageRank as a signal
With 47.33% of links on pages with a TBPR 0, that is a signal of lower quality links. The next 29.7% were on pages with PR -1 (Grey TBPR) – Usually a sign of a deindexed domain.
Coupled with the fact that another 9.31% were on pages with PR1, you can easily see that the profile looked unnatural and suggested a problem that needing investigation even before started to look at individual sites. It was a very quick way for us to eliminate the small amount of high PageRank links though!
Of course, TBPR has not been updated for a long time and may never be again – so we can never reply on this as the sole metric se look at – It is one of the many data points we gather and analyse.
Link location as a signal
We also use the location of the link to give an overview of the link profile. In the example we were investigating here we flagged a lot of footer and forum style links instantly. We did have a large % of “unknown” to work through.
The danger of forum links in an industry like this is that people discuss sites and as this was a currently visible property portal and features a large selection of properties across a large country so we expected to see an element of nature in these.
Sitewide links as a signal
Sitewide links legitimately exist on the web as bloggers give mentions on the blogroll to each other or banners announcing sponsorship or promotion of an event or entity which you wish to appear on every page. Unfortunately, SEO’s have used this to their advantage over the years and there is a good chance that if the link is not there legitimately, Google may notice it as a footprint and penalise your site based on this.
You can never take the % of the links as sitewides as a spam indication alone, but it does help when gauging the level of spam within a profile.
When you start to audit and investigate a link profile you just never know what you will find!
In this case, the domain our client was looking to buy had –
- Over 2000 unique referring domains coming from .blogspot links via a template theme created on behalf of the domain owners.
It was difficult to say how long this practice has been happening for. Many site look to have been built between 2010 and 2014. These links are coming from a variety of sites that are not relevant to the property industry, including sex and porn sites. The theme is a relatively good looking theme so is probably still being downloaded and used leaving control of these links out of our hands and with commercial anchors.
- A number of lower quality directory style links, again with what could be classed as commercial anchors.
- Looking at a top level LinkRisk report you can see the site has very few if any good or low risk links.
Where a site has even a small amount of brand visibility and/or mentions you would expect to see at least a handful of good links – especially when we are looking at 12,504 referring domains to the site! You can also see from the images above that one of the sites was a clear leader when it comes to attracting “good” links and is in a much stronger position than the other 2 although we did not investigate the rivals to the same level of depth as the site the client was looking to purchase.
- Using the SEMRush visibility graph we can see that the site our client was looking to purchase saw a spike in visibility and rankings as the current owners try to inflate the stats prior to sale.
- We can also see that visibility was a lot higher prior to May 2013 which coincides with Penguin 2.0 (#4).
The domain crowding update may have also affected the stats here as in a marketplace with a few very large players, a reduction in multiple page listings will impact traffic and visibility.
- Organic growth prior to May 2013 looks consistent with what we would expect with the visibility rising on a monthly basis.
SEO activity was massively increased Q1 2013 and it worked – for a while.
Whilst traffic for August was marginally increased vs July, it is too early in a recovery process to say whether the visibility will return – especially based on what is in the link profile and based on the fact the site owner is looking to maximise the visibility to sell the site.
As a whole the industry seems to still be reliant on older style linking techniques. The fact that the sites’ content is not English may explain the reason why Google has not yet penalised any of the main competitors.
- Low Quality Links are present in all three site’s backlink profiles, with the site our client was looking to purchase relying heavily on paid, optimised footer links and older style directory links.
This is not a sustainable technique, and the percentage of lower quality links is likely to ring alarm bells when Google eventually evaluates the site.
- A backlink cleanup for the domain would be a lengthy job, and would have involved culling the majority of the backlinks pointing to the site.
This would be the only option to help future proof the site against any unwanted manual penalties and given what would be left, you may as well start again.
- The over-optimised, commercial anchor text distribution is also a cause for concern, and once again linked to the lower quality footer links.
This blatantly goes against Google’s quality guidelines and it’s quite clear to see that this technique has only been employed for the SEO benefit.
- Further investigation is recommended into the sudden rise in traffic and visibility over the past month.
This is something that can be checked in the Site’s Google Analytics account, which will give an accurate representation of visitor numbers for the target country and audience.
From the findings of the report we were of the opinion that unnatural link building strategies have been implemented in the past to actively manipulate the search engine rankings and visibility of the site.
The link building practices employed are often used as a reason for Google to impose a manual penalty on a site, which means the site is currently in a very vulnerable position. In the event of a manual penalty being imposed, the current traffic levels could not be sustained. The majority, if not all inbound organic traffic would be lost until the penalty was removed.
Our advice to the client in this instance was to either walk away or renegotiate the price drastically.
Using our SEO skills and the link auditing and data tools we have at our disposal, we could easily identify the practises used and help the client to understand the risks involved in spending $1,000,000 on a business reliant on organic search.
Whilst this kind of work is often not cheap as it is very skilled and labour intensive, it should be seen as a crucial element of your due diligence when looking to purchase a business where the value is built on the organic visibility.
I know for sure they would rather have paid us what they did than spend $1,000,000 on a site built on sand.
- Looking to buy an online business and need advanced information on the history / visibility?
- Seen a drop in rankings on a site you have bought and cannot understand why?
- Worried after reading this and just want to know how risky your link profile is?